User needs and operational requirements for MiniFaros assistance system

Transcription

1 MINIFAROS Small or medium-scale focused research project User needs and operational requirements for assistance system Deliverable No. D3.1 Workpackage No. Task No. Coordinator Authors Status: WP3 A3.1 - A3.3 Requirements and User Needs 3.1 Accident Review and Relevant scenarios, 3.2 User needs 3.3 Sensor Requirements Kay Fuerstenberg, SICK AG Torbjörn Johansen, Volvo Technology Corporation, VTEC; Radim Hrabica, SKODA AUTO a.s., SKO; et altera Public Version No: 1.0 File Name: MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Issue Date: 15 October 2010 Project start date and duration 1 January 2010, 36 Months MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 1 of 45

2 EXECUTIVE SUMMARY The project vision is to enable accident-free traffic by the use of effective environment perception systems. Laser scanners are the predominant generic environment sensing technology. The project aims to develop a low cost novel miniature laser scanner for advanced driver assistance systems, ADAS, for vehicles. The developed laser scanner aims to be affordable, small and lightweight in order to significantly increase the penetration of ADAS in Europe. ADAS functions have been developed in several European projects during the last decade but the penetration rate is still very low on the market. One reason for this is that these functions are relatively costly. The ADAS functions have so far almost only been introduced in the full-size luxury car class segment. Most of the vehicles in Europe however are mid-size and smaller classes where these functions have not yet been introduced or demanded. A low cost, small size laser scanner would certainly alter this. To reach the objectives of small and low cost laser scanner a number of new techniques have to be developed and evaluated. The rotating mirror in laser scanners has previously been realized with a macro mechanical scanning system. This relatively large moving part in the sensor has not been fully accepted by OEMs even though it has been proved reliable. The use of a MEMS (Micro-Electro-Mechanical system) mirror in the novel laser scanner might alter this. By developing a MEMS mirror for the laser scanner will also enable downsizing of the sensor. The receiver and the Time-to-Digital-Converter (TDC) are the major integrated circuits in the sensor. These components are essential when reducing the size and cost of the sensor. The integration of these into a common circuit will be an advantage regarding cost and size. It will also enable other benefits like compensation of timing error and a possibility to measure several distances in a single laser pulse that will be beneficial for operation in bad weather conditions like rain and fog. The use of free-formed optics and aspheric surfaces developed for the laser scanner will reduce the sensor size but enabling a very large field of view. The free-form optics will reduce the number of optical components and lower production cost further. Accident review and analyses from European accident statistics have been performed and shows that pedestrian protection and pre-crash functions are the main scenarios that can be addressed by an ADAS function using data from the laser scanner. In total around 54% to 82% of all severe accidents for cars and trucks could be addressed by ADAS safety functions that could be using a laser scanner. The laser scanner can also be used for a number of ADAS comfort functions as stop and go and parking assistance. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 2 of 45

3 In this sensor development project exemplary ADAS functions already developed for safety will be adapted to the laser scanner. There have been several European projects aimed for the development of safety functions and a survey of the state-of-the-art ADAS functions have been performed in order to set the requirements on the laser scanner correctly. The requirements are described in functional terms regarding range, field of view, accuracy et cetera. The state-of-the-art ADAS survey show that the general requirements of the laser scanner sensor would be: Range: 80 meters Range accuracy: 0.1 meters in near-field and 0.3 meters else Field of view: 250 degrees Angular accuracy: 0.25 degrees Update frequency: 25 Hz Additionally more general requirements on the laser scanner are also specified because they will have to comply with automotive standards to be fitted and operated in vehicles. Object recognition algorithms have to be developed for the laser scanner for the safety applications addressed by the sensor. This includes the development of enhanced object recognition and tracking algorithms and performing improved object classification in order to be able to decide the correct strategy for avoiding objects in the traffic environment. The laser scanner will be shown and demonstrated serving various safety applications in vehicle environment. Both a car and a truck demonstrator will be used. To show the huge potential of the laser scanner in non automotive applications infrastructure based perception will be developed and demonstrated. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 3 of 45

4 Revision Log Version Date Reason Name and Company Document structure Torbjörn Johansen, VTEC First ICCS input Vassilis Kaffes, ICCS First SKO input Radim Hrabica, SKO 0, Compilation of partner input and first version. Torbjörn Johansen, VTEC Input from SICK Kay Fuerstenberg, SICK Second ICCS input Vassilis Kaffes, ICCS Second SKO input Radim Hrabica, SKO Compilation of partner input and second version Torbjörn Johansen, VTEC Partner input Radim Hrabica, SKO Vassilis Kaffes, ICCS Kay Fuerstenberg, SICK Final document for review Torbjörn Johansen, VTEC Peer review report and document comment Vassilis Kaffes, ICCS Radim Hrabica, SKO Review input and document comment Florian Ahlers, SICK Final document proposal Torbjörn Johansen, VTEC 0.42 July, August Partner input and peer review reports Compilation of partner input to final document Florian Ahlers, SICK Radim Hrabica, SKO Torbjörn Johansen, VTEC Peer review report Tapani Mäkinen,VTT Final document proposal Torbjörn Johansen, VTEC Partner comments Axel Jahn, SICK Kay Fuerstenberg, SICK Final document proposal Torbjörn Johansen, VTEC Partner input Kay Fuerstenberg, SICK Final document Torbjörn Johansen, VTEC Final layout Kay Fuerstenberg, SICK MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 4 of 45

5 Table of contents EXECUTIVE SUMMARY... 2 Revision Log... 4 Table of contents... 5 List of Figures... 6 List of Tables Introduction Advanced driver assistance systems overview Accident review and relevant scenarios Databases investigated Main results of accident review Accidents by car segments Accidents by road users Accident review for cars Used statistical sample Types of accidents for cars Addressed accident types by laser scanner for cars Accident review for trucks Types of accidents for trucks Addressed accident types by laser scanner for trucks Accident review for pedestrians Accident type for pedestrians Time of day and road conditions Addressed accident types by laser scanner for pedestrians Accident review for frontal crash Addressed accident types by laser scanner for frontal crashes User needs Requirements from the road user Demographical data for survey on ADAS Results from road user survey on ADAS Sensor requirements Functional requirements Sensor output requirements Processing output requirements Range requirements - object recognition Range accuracy Field of view Functional requirements for Infrastructure laser scanner Object detections and measurement Estimation of Needed Coverage Area Data acquisition rate Practical Limitations Non-functional requirements General requirements Operational and environmental conditions Material requirements Vibration requirements Requirements for wiring and contact pins Electrical requirements Conclusions References MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 5 of 45

6 List of Figures Figure 1. Approach for deriving laser scanner requirements... 7 Figure 2. Road safety evolution in EU, [7] Figure 3. Accidents distributed by car segments, [11] Figure 4. Road fatalities by type of user in 2008, [7] Figure 5. Distribution of fatalities by mode of transport, [8] Figure 6. Statistical sample of age, [11] Figure 7. Distribution of accidents by type, [11] Figure 8. Accident types of serious or fatal accidents for heavy trucks, [12] and [13] Figure 9. Initial speed of vehicles for pedestrian collisions, [11] Figure 10. Initial speed of vehicles for frontal collisions, [11] Figure 11. ADAS user interest in targeted size vehicle segments [14] Figure 12. Comparison of market price and customers ideas of price, [14] List of Tables Table 1: Advanced driver assistance functions, ADAS, overview Table 2: Accident types for cars. Possible driver assistance systems and potential for laser scanner Table 3: Truck manoeuvre in accidents, [10] and [9] Table 4: Accident types for trucks. Possible driver assistance systems and potential for laser scanner Table 5. Distribution of pedestrian accidents, [11] Table 6. Distribution of frontal crashes, [11] Table 7: Approximate sensor requirements for ADAS, [2] Table 8. Wide band random vibration profile values for body mounted parts Table 9. Connector wiring MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 6 of 45

7 1. Introduction The project Low cost miniature laser scanner for environmental perception,, is a sensor development project aimed to significantly increase the penetration of advanced driver assistance systems, ADAS, on the automotive market. The project vision is to have a accident-free traffic environment by the use of effective environment perception systems. Laser scanners are the predominant generic environment sensing technology. Poor human perception and assessment of traffic situations stands for the largest amount of traffic accident with fatal or severe injury outcome. Several safety functions are developed in order to prevent or mitigate many of these accidents. The system cost for these functions are often relatively high however so vehicles are rarely equipped with these systems, especially when it comes to smaller cars or commercial vehicles. In order to develop the laser scanner it has to be defined how it should be used in order to set adequate requirements on it. This deliverable describes the relevant scenarios and operational requirements for the laser scanner for ADAS functions. The sensor requirements will be derived from the top-down approach from accident scenarios to requirements according to Figure 1. Accidentology Relevant scenarios Preliminary laser scanner performance State-of-the-art ADAS functions Object recognition requirements User needs laser scanner requirements Figure 1. Approach for deriving laser scanner requirements. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 7 of 45

8 The first task will be to perform accident analysis and accident review in order to find the most relevant scenarios where accidents can be avoided or mitigated. Statistics from European countries and previous ADAS Integrated Projects have been studied. The result for this task is presented in Chapter 2 Accident review and relevant scenarios. These general statistics can help us to select the most useful functions to reduce accidents to be covered by the laser scanner. There will be no safety function development in this project but the laser scanner will be used with current state-of-the-art functions developed in previous ADAS projects. These functions will be summarised in the next subsection briefly. User needs on the sensor system is described in Chapter 3 User needs. Here are mainly the end user needs for an increased use of ADAS functions reviewed. Automotive industrial user needs and sensor requirements are described in chapter 4. They are derived from the relevant scenarios, their safety functions and the user needs described in chapter 2 and 3. In order to increase market penetration especially in the class of small vehicles, the market price has to be very low. The laser scanner is also developed to be small and lightweight in order to enable an easy integration in the vehicle design. The aim with the project is to develop a laser scanner that in series production will be relatively cheap to comparable sensors with other technologies such as radars. To reach the objectives of small and low cost laser scanner a number of new techniques have to be developed and evaluated. The rotating mirror in laser scanners has previously been realized with a macro mechanical scanning system. This relatively large moving part in the sensor has not been fully accepted by OEMs even though it has been proved reliable. By the use of a MEMS (Micro-Electro-Mechanical system) mirror in the novel laser scanner might alter this. By developing a MEMS mirror for the laser scanner will also enable downsizing of the sensor. The receiver and the Time-to-Digital-Converter (TDC) are the major integrated circuits in the sensor. These components are essential when reducing the size and cost of the sensor. By integrating these into a common circuit will be an advantage regarding cost and size. The integration will also enable other benefits like compensation of timing error and a possibility to measure several pulses in a single event that will be beneficial for operation in bad weather conditions like rain and fog. The use of free-formed optics and aspheric surfaces developed for the laser scanner will reduce the number of optical components and also make it possible to integrate the optics and the mechanics in the sensor. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 8 of 45

9 This will also make it possible to decrease the sensor size and lower production cost further Advanced driver assistance systems overview This project s goal is to develop a laser scanner able to serve several ADAS functions. The driver assistance systems already on the market or at research level are described shortly in Table 1. The functions are developed by previous European thematic network ADASE, Advanced Driver Assistance Systems in Europe, [1] and Integrated Project PReVENT [2] with sub-projects. These systems are using various sensors to monitor traffic environment. Often multiple sensors are used and their data is fused on a higher or lower level. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 9 of 45

10 Table 1: Advanced driver assistance functions, ADAS, overview. Function acronym CAS FCW Pedestrian protection LCW LDW LKA LCA Start inhibit Intersection assistance CMS AEB Pre-crash DAS ACC S&G Parking assistant Meaning of acronym Collision Avoidance Systems Forward Collision Warning Lateral Collision Warning Lane Departure Warning Lane Keeping Assistant Lane Change Assistant Collision Mitigation Systems Automatic Emergency Braking Driver Assistance & Comfort Systems Adaptive Cruise Control Stop and Go assistance Short system description Warn for obstacles including other vehicles in front of the host vehicle. Warn and act if pedestrians are in front of the host vehicle. Warn for vehicles coming from the side of the host vehicle and crossing the path. Warn if the host vehicle is about to unintentionally exit the current lane or road. Functionality like the LDW but with ability to actively steer back. Warn if the adjacent lane is occupied if a lane change manoeuvre is initialised. Inhibit host vehicle (truck) to start when an object is in the blind spot in front. Monitor and warn for oncoming traffic in intersection scenarios. System that automatically start to brake if a frontal collision is unavoidable. System that prepare passive safety systems like airbag, seatbelt pre-tensions, active hoods for pedestrian protection, etc, when a collision is unavoidable. System that automatically keeps a safe distance to the vehicle in front. Like ACC but for lower speeds and ability to stop host vehicle and automatically start again while driving in queue. Monitoring objects close to host vehicle in parking situations. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 10 of 45

11 2. Accident review and relevant scenarios Initially an accident analysis has been performed to survey the importance of the various scenarios that will be relevant for the laser scanner as well as the importance of other parameters, in order to derive the selection of use cases. A number of data sources have thoroughly been investigated leading to frequency percentages per accident type. Finally, scenarios relevant to laser scanner have been identified and scenarios addressed for further development of the sensor have been described Databases investigated The sources investigated for the accident analysis are the following: Statistic results from literature for the following two European databases : Community Road Accident Database, CARE, [7], European Commission, 2006 Annual Reported statistics for Road Casualties in Great Britain, [8], 2008 Results from the Integrated European Project PReVENT, [2] National Highway Traffic Safety Administration, NHTSA, General Estimates System, GES [9], 2008 Fatality Analysis Reporting System, FARS [10], 2008 German In-Depth Accident Survey, GIDAS [11] 2.2. Main results of accident review Looking at the accidental data in Europe in the last two decades the overall number of road accidents involving personal injuries has only decreased slightly. However the number of road fatalities was reduced by almost 50% as shown in Figure 2. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 11 of 45

12 Figure 2. Road safety evolution in EU, [7]. Additionally the injured and accidents curves displayed in Figure 2 have almost the same shape. These facts can be interpreted as even though the traffic in terms of fatal accidents improved significantly, the traffic facilities are still not safe enough to increase injuries prevention. Not to be mistaken, however, the number of of deaths in 2008 is still too many to be satisfied with. More precautions are to be taken by all the traffic stakeholders in Europe. The economical impact is also not to be underestimated. For example only in Czech Republic the total costs of accidents has almost achieved 306 billion in 2008, which was approximately 11% of the republic s annual budgetary deficit. It is important to realise that considering the growth of quantity of cars in Europe with over 40% since 1990 it would be false to declare that traffic safety is not improving. In terms of life saving and injury prevention, the steps taken in Europe, with e.g. mandatory usage of systems like ESP or passive pedestrian protection, are definitely helping. But there is a belief that the safety can be improved even more radically by increasing the ADAS penetration in European automotive market. For this we need cheap sensor systems. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 12 of 45

13 Accidents by car segments The need of developing cheap sensors suitable for small cars is displayed in Figure 3. The figure shows in what segment the car, which belongs to the driver responsible for an accident, fall in. Notice that 56% of all accidents are caused by cars belonging to segments A, B and C. These segments are unfortunately literally untouched by advanced driver assistance and active safety systems today. Even the penetration of D segment by these systems is very low. Figure 3. Accidents distributed by car segments, [11]. Examples of cars in the different segments in Figure 3 are. Subcompact: Opel Corsa, VW Polo, Ford Fiesta, Skoda Fabia, Audi A2, Fiat Punto, Ford Fusion, Honda Jazz Compact: VW Golf, Opel Astra, Ford Focus, Mercedes A-Klasse, Audi A3, Skoda Octavia, Toyota Corolla Mid size: Audi A4, VW Passat, Opel Signum, Ford Mondeo, Honda Accord, Mazda 6, Škoda Superb, Volvo S40/S60 Full size: Audi A6, Mercedes E-Klasse, Volvo V70, Jaguar S-Type, Mercedes CLK, Peugeot 607, Saab 9-5 Full size luxury: Audi A8, BMW 7er, Mercedes S-Klasse / CLS, VW Phaeton Sport cars: BMW 6 Series, Mercedes CLK, Volvo C70, Volkswagen Eos, Audi TT, BMW Z4, Porsche Boxster/911 MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 13 of 45

14 Accidents by road users The distribution of fatalities among the various road users is depicted in Figure 4. Drivers constitute the highest number of fatalities with about 60%, followed by vulnerable road users, VRU, like pedestrians, 20%, passengers 19% and others/not specified 1%. Figure 4. Road fatalities by type of user in 2008, [7]. According to the annual statistical report from Great Britain, [8], the percentage of passenger car driver fatalities are 53% and pedestrians 19%, see Figure 5. Car occupants and pedestrians account for the vast majority of road fatalities. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 14 of 45

15 Figure 5. Distribution of fatalities by mode of transport, [8] Accident review for cars For a more comprehensive accident analysis and scenario selection for passenger car more than general statistical data is required. For this purpose the data obtained from GIDAS, [11], is used Used statistical sample The selected statistical sample of the GIDAS, [11], database in this survey is: Accidents recorded between 07/1999 and 02/2010 Accidents between passenger car and any other traffic participant No selection on severity set for accidents with injuries of any kinds Accidents caused by breaking the law (e.g. traffic light violation) is included The statistical sample consists of accidents with the attributes seen in Figure 6 below. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 15 of 45

16 Figure 6. Statistical sample of age, [11] Types of accidents for cars The accident data obtained from [11] allows categorization of accidents into the following basic groups shown in Figure 7. Figure 7. Distribution of accidents by type, [11]. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 16 of 45

17 1. Start, stop: Collision with another vehicle which starts stops or is stationary. Starting or stopping are here to be seen in connection with a deliberate stopover which is not caused by the traffic situation. Stationary vehicles within the meaning of this kind of accident are vehicles which stop or park at the edge of a carriageway, on shoulders, on marked parking places directly at the edge of a carriageway, on footpaths or parking sites. The traffic to or from parking spaces with a separate driveway belongs to No. 5 kind of accidents. 2. Drive ahead, wait: Collision with another vehicle moving ahead or waiting. Accidents caused by a rear-end collision with a vehicle which either was still moving or stopping due to the traffic situation. Rear-end collisions with starting or stopping vehicles belong to the No. 1 kind of accidents. 3. Lateral, same direction: Collision with another vehicle moving laterally in the same direction. Accidents occurring when driving side by side (sideswipe) or when changing lanes (cutting in on someone). 4. Drive towards: Collision with another oncoming vehicle. Collisions with oncoming traffic, none of the colliding partners having had the intention to turn and cross over the opposite lane. 5. Incurve, cross: Collision with another vehicle which turns into or crosses a road. This kind of accident includes collisions with crossing vehicles and with vehicles which are about to enter or leave from/to other roads, paths or premises. A rear-end collision with vehicles waiting to turn belongs to the No. 2 kind of accidents. 6. Vehicle-pedestrian: Collision between vehicle and pedestrian. Persons who work on the carriageway or still are in close connection with a vehicle, such as road workers, police officers directing the traffic, or vehicle occupants who got out of a broken down car are not considered to be pedestrians. Collisions with these persons are recorded under the No. 10 kind of accidents. 7. Obstacle at roadbed: Collision with an obstacle in the carriageway. These obstacles include for instance fallen trees, stones, lost freight as well as unleashed animals or game. Collisions with leashed animals or riders belong to the No. 10 kind of accidents. 8. Run off to the right 9. Run off to the left: Leaving the carriageway to the right or left. These kinds of accidents do not involve a collision with other road users. There may however be further parties involved in the accident, e.g. when the vehicle involved in the accident veered off the road trying to avoid another road user and did not hit him. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 17 of 45

18 10. Other traffic accidents: Accident of another kind. This category covers all accidents which cannot be allocated to one of the kinds of accidents listed under Nos. 1 to Addressed accident types by laser scanner for cars Figure 7 shows that there can be four categories of accidents identified which represent the most of the accidents. Frontal crash related accidents can be considered types 1, 2 and 4 with a total share of 32%. These accidents can be avoided or at least partially mitigated by a pre-crash system. Therefore pre-crash is selected as one of the systems determined to be developed for demonstrator passenger car. The second group of accidents is related to leaving or more precisely running off the roadbed. These are the categories 8 and 9 with a total share of total 24% of the accidents. To handle these types of accidents, functions similar to lane departure warning, LDW, are required. The laser scanner will not be able to be used for lane monitoring. These types of accidents are therefore beyond the scope of this project. The third group is somehow related to intersections and junctions. To this group belongs the category 5 of accidents displayed in Figure 7 with approximately 20% of the accidents. The fourth group of accidents is with pedestrians. With a total percentage of 18% of all accidents, the need of using pedestrian protection system for the demonstrator passenger car is obvious. The scenario types in Figure 7 can be summarised in Table 2 where also possible assistance systems for the accidents are displayed. In the table potential for the laser scanner in relevant assistance systems are also shown based on preliminary performance. In total about 54%- 74% of the accidents can be addressed with an ADAS function that get data from a laser scanner. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 18 of 45

19 Table 2: Accident types for cars. Possible driver assistance systems and potential for laser scanner. Accident type Description Longitudinal collision Longitudinal collision Lateral collision Longitudinal collision Intersection collision Pedestrian collision Longitudinal collision Lane departure Lane departure Misc. collisions Frequency 3% 19% Possible ADAS function FCW, Pre-crash, AEB FCW, Pre-crash, AEB, ACC, S&G laser scanner potential Yes Yes 3% LCW, LCA Yes 10% 20% 18% 1% FCW, Pre-crash, AEB Intersection assist Pedestrian protection FCW, Pre-crash, AEB Yes Yes for urban situations Yes Yes 14% LDW, LKA No 10% LDW, LKA No 2% - No 100% 54%-74% 2.4. Accident review for trucks In this paragraph there will be a summary of the data found in different databases, regarding accidents involving heavy trucks Types of accidents for trucks In order to find relevant scenarios for the laser scanner a more detailed analysis of accidents with heavy vehicles have to be evaluated. Analysis from [12] and investigations by [13] has been used for this review. Figure 8 shows the distribution of different accident types from these studies. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 19 of 45

20 Figure 8. Accident types of serious or fatal accidents for heavy trucks, [12] and [13]. The types of accident shown in Figure 8 can be described as: 1. Single, run off. This accident type is the type where most heavy vehicle users are killed or seriously injured. The accident is often in rural areas or highway approaches. 2. Drive ahead, wait: Collision with another vehicle moving ahead or waiting in the same lane. 3. Drive towards. Collision with an oncoming vehicle. 4. Incurve, cross. Collision with another vehicle which turns into or crosses the road. Intersection accidents. 5. Lateral, same direction. Collision with another vehicle during lane changing manoeuvres or while overtaking or being overtaken. 6. Vulnerable road users. Collision between vehicle and pedestrian, cyclist, motorcyclist or other vulnerable road user. 7. Other accidents. These accidents are other not suitable in the above types. It could for example be other vehicles running into the heavy truck rear end while stationary. Statistics from [9] and [10] shows that large trucks accounted for 8% of the vehicles in fatal crashes, but only 2% of the vehicles involved in injury crashes and 4% of the vehicles involved in property-damage-only crashes. Of the 4,066 large trucks involved in fatal crashes, 74% were combination trucks. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 20 of 45

21 Vehicle manoeuvres of trucks involved in an accident are shown in Table 3. In a majority of accidents the truck is going straight when the collision occur indicating head on collisions. This manoeuvre involves then accidents with vulnerable road users. The second largest is the single accident type when a curve is negotiated and control over vehicle is lost. After that vehicle stopping in lane and intersection manoeuvres while making a left turn is the most common. Table 3: Truck manoeuvre in accidents, [10] and [9]. Share of fatal Vehicle manoeuvre accidents, [10] Share of all accidents, [9] Going Straight 69.5% 53.2% Decelerating in Traffic Lane 2.8% 5.3% Accelerating in traffic lane 0% 0.2% Starting in Traffic Lane 0.8% 2.0% Stopped in Traffic Lane 5.7% 11.9% Passing or Overtaking Another Vehicle 1.1% 0.9% Disabled or Parked in Travel Lane 0.02% 0.14% Leaving a Parked Position 0.07% 0.3% Entering a Parked Position 0.05% 0.1% Turning Right 1.8% 3.3% Turning Left 4.5% 11.2% Making U-Turn 0.4% 0.5% Backing up (not parking) 0.9% 0.7% Changing Lanes or Merging 1.7% 3.2% Negotiate a Curve 8.2% 5.0% Other/unknown 0.8% 1.9% Addressed accident types by laser scanner for trucks The heavy truck accident analysis shows that there are three categories of accidents that represent the most of the accidents as shown in Figure 8 in section Front crash related accidents are the major accident types and covering in total 37% of all serious or fatal accidents. That is Type 2 and 3 in Figure 8. To decrease these accidents or to decrease the severity of the accidents forward collision warning systems, FCW, automatic emergency braking, AEB, or other pre-crash functions could be of use. The pre-crash functions could be preparing the vehicle for a crash like deploying airbags, MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 21 of 45

22 pretension of seatbelts when the accident is unavoidable. Functions for this are developed in [4]. Other functions that could be of help are adaptive cruise control, ACC, and Stop and Go, S&G. All these functions could be a possible for the laser scanner. A second accident type is the collision with vulnerable road users, VRU, with 24% of the accidents. These accidents often occur due to a limited vision for the driver due to large blind spot areas of a truck. By monitoring the environment close to the truck by sensors the number of this type of accidents could be decreased. Functions for VRU monitoring and warnings like blind spot detection could also be possible for the laser scanner. The third main accident type is conflicts at intersections and crossing traffic with 18% of the accidents. The scenario types in Figure 8 are summarised in Table 4 where also possible assistance systems for the accidents are displayed. In the table the laser scanner potential for relevant ADAS functions are also shown based on preliminary performance of the laser scanner. In total about 64% to 82% of the accidents can be addressed with an assistance system that might incorporate a laser scanner. Table 4: Accident types for trucks. Possible driver assistance systems and potential for laser scanner. Accident type Description Lane departure, roll-over Longitudinal collision Longitudinal collision Intersection collision Lateral collision VRU collision, Misc. collisions Frequency Possible ADAS function laser scanner potential 5% LDW, LKA No 8% 29% 18% FCW, Pre-crash, AEB, ACC, S&G FCW, Pre-crash, AEB Intersection assist Yes Yes Yes for urban situations 3% LCW, LCA Yes 24% Pedestrian protection, FCW, Pre-crash, AEB Yes 13% - No 100% 64%-82% MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 22 of 45

23 2.5. Accident review for pedestrians As shown in the previous chapter pedestrians are an exposed road user for accidents in Europe. In this paragraph a summary of the outcomes elaborated from different database investigations on accidents with pedestrians is presented. Pedestrian and other vulnerable road user fatalities are more common within urban areas than outside since the pedestrian density is higher in cities than outside Accident type for pedestrians The most common collision opponents for pedestrians are passenger cars. Indeed 77% of pedestrian fatalities in 2008 in Great Britain, [8], are related to passenger cars as collision opponent. An analysis in APALACI, [4], a subproject of the PReVENT IP, [2], proved that concerning the pedestrian movement, 65% of involved pedestrians have crossed the road from the right to the left, for right hand traffic. About 60% of the pedestrians were walking and 20% were running at the time of collision. The remaining 20% were either stationary or moving in some other way. Other results from [2] show that about 70% of the pedestrian collisions the impact occurs at the front end of the vehicle. The point of impact in the front region is rather evenly distributed between front 40%, right 35% and left side 25%. According to GIDAS, [11], 97% of accidents with pedestrians occur in urban areas and three main groups of pedestrian accidents can be identified as depicted in Table 5. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 23 of 45

24 Table 5. Distribution of pedestrian accidents, [11]. Accident type I. Case vehicle turning, pedestrian crossing road II. Case vehicle driving straight, pedestrian crossing road Case description Turn left or right, pedestrian coming either from left or right, including reduced visibility cases a. Pedestrian crossing road from the left or right, visibility was not reduced Pictogram examples Ratio to pedestrian accidents [11] All Severe or fatal 7.5% 4.5% 55.4% 79.8% b. Pedestrian crossing road from the left or right, visibility was reduced 24.8% 9.0% III. Case vehicle driving straight, pedestrian not crossing road IV. Special cases Pedestrian moving longitudinally, same or opposite direction to CV e.g. pedestrian suddenly appears on the roadbed or is seen in the last moment 2.6% 6.7% 1.7% 0% 92% 100% As can be seen in Figure 9 that is linked to Table 5 the Type I of the accidents happen at lower initial speeds. In fact 80% of all accidents take place up to a case vehicle speed of 30km/h. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 24 of 45

25 Figure 9. Initial speed of vehicles for pedestrian collisions, [11] Time of day and road conditions Regarding environmental factors the investigation in [2] and subproject [4] identified that most of the accidents, 68%, happened during daylight hours. A large portion of the remainder, 17%, occurred on unlit roads at night and further 8% happened on dark roads with streetlights. Visibility was reduced by fog, mist, or heavy rain in 7% of the cases. 67% of the accidents occurred on dry roads, 31% on wet roads, and 2% on icy, slippery, or snow covered roads Addressed accident types by laser scanner for pedestrians The accident analysis of pedestrian summarised in Table 5 shows that covering accident when pedestrians cross the road and the vehicle is either is driving straight or is making a turn (types I and II) would prevent or mitigate consequences of 88% of pedestrian related accidents. It is difficult to cover all cases however. For example it may prove impossible to prevent accident where the pedestrian suddenly appears on the road. Not only to mention accident type IV, but also Type IIb, where the visibility is reduced by an obstacle. Nevertheless, even when not considering these scenarios, more than 65% of pedestrian accidents could be covered. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 25 of 45

26 2.6. Accident review for frontal crash The other major accident type is frontal crash and it is studied more in detail in this section. The majority of all accidents happen with at least one vehicle hitting frontally. This vehicle is considered to be the case vehicle, CV. The selected scenarios describe the most frequent types of frontal crash accidents. Accidents happening at junctions when one vehicle is turning and crashes into second vehicle driving straight are not considered here. The distribution of different accident types for frontal crash is shown in Table 6. Table 6. Distribution of frontal crashes, [11]. Accident type Case description Pictogram examples [11] All Ratio to frontal accidents severe or fatal I. CV hits frontally to another vehicle Initial speed before the crash is a sum of both vehicles initial speeds. 22.6% 68.4% II. CV driving straight, another vehicle changing to same lane or is already moving in the same lane III. CV driving straight, another vehicle is moving with minimal longitudinal speed (e.g. turning, driving out of parking place etc.) Initial speed before the crash is difference of both vehicles speeds Initial speed is practically speed of the CV 33.0% 23.7% 39.2% 5.3% IV. CV hits into solid obstacle 5.1% 2.6% 99.9% 100% MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 26 of 45

27 The initial speed of the case vehicle, CV, in the accident types in Table 6 are shown in Figure 10. Figure 10. Initial speed of vehicles for frontal collisions, [11]. As can be seen in Figure 10, over 80% of accidents happen in range up to 40 km/h. Type II of accidents is the exception. In this type 70% of accidents happen in this range. It s important to realize that in Type I and Type II of accidents the initial speed of CV is not the actual crash speed. In Type I when two cars are hitting frontally, the crash speed is a sum of both vehicles speeds. In Type II it is on the contrary the difference of these speeds. Actual relative speeds of accidents are not available and depend on how much the drivers are able to brake prior to the impact Addressed accident types by laser scanner for frontal crashes Accident types for frontal crashes have been discussed in the accident reviews for cars and trucks. About a third of all severe accidents are related to frontal crash types. Around 80% of the accidents are at speeds in the range of 0-50 km/h according to Figure 10. The most severe accidents are head to head according to Table 6, Type I. This accident type is to be addressed by the laser scanner with a Pre-crash or FCW functionality. Type II in Table 6 is also addressed by these functionalities. Other functions to keep a safe distance or S&G or ACC are also addressing this accident type. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 27 of 45

28 3. User needs Considering the user needs on the laser scanner, several needs can be addressed. These will be allocated into two basic user areas and described in following parts. Needs of the road user this is the first thing needed to be taken into account. Here, the road user is actually the customer, who will purchase the system at the end. Its demands are mainly oriented on price to function ratio. The needs from the industrial user the developed sensor has to fulfil certain characteristics to be easily adopted into automotive sector. These characteristics include electrical and mechanical properties, which will be described in the sensor requirement in Chapter Requirements from the road user One of the main purposes of the newly developed sensor is to increase the volumes of cars equipped with advanced driver assistance systems, ADAS in order to significantly reduce the number of accidents on the roads of Europe. This goal is only achievable through finding an optimal ratio between the customer s satisfaction and producer s business intention. To find such a ratio is quite difficult however, especially in low class and mid class vehicle segments. According to AUTOTECH, survey in 2006, [14], which was done in scope of people from United Kingdom, Germany, Italy, France and Spain, there are significant differences between what owners of different vehicles think to be a reasonable price to even consider the purchase of ADAS. To correctly address the best ADAS functions, which are to be focused on, confronting the market analysis with the accident statistics is highly contributory Demographical data for survey on ADAS Detailed demographical data are to be found in [14]. Only overall basic data shall be described here: In the survey 57% were men and 43% women. 65% of respondents were married or living with partner, 25% were singles and never married. Remaining percentage of respondents was in other marital status. 9% of respondents were in age of 18-24, 22% of respondents were in age of 25-34, 23% of respondents were in age of 35-44, 22% of respondents were in age of and 21% of respondents were in age of % of respondents didn t have children in household. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 28 of 45

29 Results from road user survey on ADAS The following figures demonstrate how people did react on presented ADAS functions before and after the actual market price was introduced to them. The actual market price in 2006 and the price considered by the users as a good value is part of the Figure 12. From above mentioned figures and by considering only the functions suitable for laser scanner, several conclusions can be made. For instance, let s have a closer look to ACC system. According to Figure 11, there was a great initial interest in ACC technology even in the segment of small size vehicles. Unfortunately, after the market price was presented to the persons questioned, they initial interest of 22% dropped to 11%. Figure 12 shows the reason for it. The good value estimated by the potential customers not even reached the half of the market price. Meaning that for small size vehicle market penetration of 22% by the ACC system, the market price would have to drop approximately to 500. Unfortunately, this would not even cover the car producer s expenses for only buying and implementing the required sensor technology into vehicles in It is also needed to be noted, that the good value price in Figure 12 was estimated by the examined, who were willing to purchase the system even before the market price was presented to them. According to results of the AUTOTECH survey, the good value price estimated by the questioned persons, who were not willing to purchase the system before the market price introduction is for about 100 lower. This makes the attempts to penetrate small size vehicle segment by ACC system even more difficult. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 29 of 45

30 Figure 11. ADAS user interest in targeted size vehicle segments [14]. Attempting to identify the best ADAS functions suitable for a cheap laser scanner, two of them is showing a greater potential than the others. According to Figure 11, the initial value of pedestrian protection system and pre-crash/radar enabled collision warning system increased even after the market price was introduced to questioned persons. With the cheap technology of environment sensing, the estimated market price addressed in AUTOTECH survey could be achievable and thus it seems that the customer s satisfaction and producer s business intention would meet in these two functions. Pedestrian protection system and pre-crash system were also identified in chapter The third ADAS function for laser scanner is to be defined for passenger car and truck demonstrator. ACC system is not to be considered for its technical complexity and wide connections to another car systems not usually included in small size cars. For lane departure aid, video camera might be needed which makes the system also not considerable. As parking aid is not primarily an active safety system, the only left to consider is blind spot detection & warning. Even though this MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 30 of 45

31 would be a reasonable choice with respect to accident statistics, two other sensors might be needed for the function to integrate (driver and passenger sides). Because we have already selected two systems which will use only one laser scanner mounted in front of the vehicle, it speaks for itself to integrate another function requiring only one front laser sensor. The example of such system is turning assistance already described in the INTERSAFE project, [5]. The other option is warning against not keeping the minimum recommended distance between vehicles. Regarding police records of Czech Republic 22.7% of all accidents were caused because of this reason. Figure 12. Comparison of market price and customers ideas of price, [14]. Note: There is a slight pricing inaccuracy in the survey of the blind spot detection & warning system and lane departure warning and aid (both marked with * in the Figure 12). Even in 2006 the price of these systems did not reach the value presented by the survey (900 and 800 respectively). The charts already calculate with corrected prices of 550 and 500 respectively, which were the common prices in 2006 in luxury vehicle segment. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 31 of 45

32 4. Sensor requirements The laser scanner needs to be prepared for use in automotive industry. Fulfilling the basic requirements described in the following chapters should prepare it for easier integration and use in possible series production later Functional requirements The ADAS functions described briefly in Table 1 and referenced to in the accident review and relevant scenarios are using various sensors to monitor traffic environment. Often multiple sensors are used. No new ADAS functions will be developed in this project specific for the laser scanner. A survey on the approximate sensor requirements for the different state-of-the-art ADAS functions and the sensor technology used today, in production or in research projects are shown in Table 7. The requirements are mainly from the PReVENT [2] sub-projects, [3], [4], [5] and [6]. In the table there is also a column for the laser scanner potential to be used based on preliminary theoretical sensor performance stated in the Technical Annex Sensor output requirements The sensor output known as sensor observation should include the following. Object track parameters such as: o ID: Provides a unique identification name for every physical property which was measured by the sensor and is of our interest. The maintenance or on time update of the IDs during the tracking process is important for the successful sense of the surrounded environment. o Position: This is the spatial location of the measured physical property. o Measurement: This is the value of the physical property as measured by the sensor element. o Confidence: This is a generic term referring to many different types of errors in measurement such as measurement errors, calibration errors or environmental errors. The errors that are not defined in sensor data sheet may be calculated internally for measurement validation by the sensor. o Time stamp: This is the time when the physical property was measured. In real time systems, the time of a measurement is often as important as the value itself. o Velocity: This is the estimated velocity of the physical property that is calculated by object tracking. o Tracking life time: This is the time a tracked object is maintained in successive scans. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 32 of 45

33 o Hidden status: Value that indicates whether the track has been associated with a measurement in the current scan and equals to zero or if it is just an estimation based on previous estimation. Object represented by contour. Contour-based object tracking requires object detection only once. Object classification. The recognized objects should be classified as: o Pedestrian o Car o Track o Motorcycle o Cyclist o Unclassified ( for other objects such as stationary objects) Processing output requirements Correct output from the processing is fundamental for a successful object detection and tracking. It must implement and fulfil the following requirements: ID maintenance: It is important to maintain the same ID for a physical property scanned by the laser during the process. Shape changes of a detected object must be overcome. Track splitting and merging: It is important to detect when tracks merge or split because it is a common reason for "ghost" track appearance or track disappearance respectively. If a track splits, then there will be two segments overlapping the track and when two tracks merge, then there will be two tracks overlapping one segment. Real or near-real time data distribution: In a safety application as applied in, the fast processing and delivery of information is just as important as the information itself. So during the design and implementation, issues regarding data availability should be of the highest priority. As a result, all the information provided to the perception modules should be transmitted with the lowest latency possible. With the term near real time it is assumed that the delay in information distribution is acceptable as long as it stays under a timing threshold which is defined by the application requirements and the algorithms capabilities. Update frequency: The update frequency is the frequency of which the tracking of an object should be updated. Sensors for detection of vehicles, VRUs and roadside objects shall have a detection rate of at least 5 Hz. Specific update frequencies are specified in Table 7 for different ADAS functions. MINIFAROS_D3.1_Requirements_and_user_needs_v10.doc Page 33 of 45