Safety Benefits of Stability Control Systems for Tractor-Semitrailers using Hardware-in-the-Loop Simulation Sangjun Park Center for Sustainable Mobility, Virginia Tech Transportation Institute 3500 Transportation Research Plaza, Blacksburg, VA 24061 Phone: (540) 231-2647 Fax: (540) 231-1555 sangjun@vt.edu Kelly Donoughe Transportation Solution Division, SAIC Phone: (703) 676-2430 Kelly.m.donoughe@saic.com Hesham A. Rakha (Corresponding author) Charles E. Via, Jr. Department of Civil and Environmental Engineering 3500 Transportation Research Plaza, Blacksburg, VA 24061 Phone: (540) 231-1505 Fax: (540) 231-1555 hrakha@vt.edu Total word count: 4911 (text) + 3,000 (12 tables & figures) = 7911 Paper 12-0791
Park, Donoughe and Rakha 2 ABSTRACT The research presented in this paper estimates the safety benefits of electronic stability control systems for tractor-semitrailers using Hardware-in-the-Loop (HiL) simulation. The HiL system used in this study consists of a pneumatic tractor-trailer braking system and a truck simulation engine, TruckSim. Additionally, the Bendix electronic stability control (ESC) system is integrated via the HiL system for the evaluation. The study evaluates the performance of the stability control system using the HiL platform for a portion of the rollover and loss-of-control crashes selected from the Large Truck Crash Causation Study (LTCCS) database. Subsequently, the evaluated performance is combined with the effectiveness rated by the UMTRI expert panel to calculate the total crash prevention ratios. This study finally estimates the crash reductions and cost benefits and provides our lessons-learned to researchers and practitioners through the course of the evaluation procedures.
Park, Donoughe and Rakha 3 INTRODUCTION Though the trucking industry accounts for 4 percent of registered vehicles and 7 percent of the total vehicle miles traveled, professional truck drivers are disproportionally involved in 8 percent of fatal crashes and 4 percent of personal injury and property-damage-only crashes (1). According to the Trucks Involved in Fatal Accidents (TIFA) database for the years 1999 to 2005, 13 percent of the fatal crashes involving heavy trucks were caused by rollovers. Research done by Wang and Council determined that there are approximately 4,400 to 5,000 truck rollovers on interchange ramps every year, with an associated cost of $405 to $460 million annually (2). Research regarding truck rollover prevention and mitigation began in the early 1990s: warning signs, rollover warning devices, and rollover training devices were studied and used for rollover prevention (3, 4). Electronic stability systems are the most recently developed technologies. These systems are multipurpose: they monitor the truck s motion, and they actively engage brakes to mitigate rolling over or jack-knifing. In the early stages, test-track testing under controlled driving and environmental conditions was widely used for the evaluation of truck safety technologies. Field operational tests (FOT) have also been a popular method because the truck s performance can be monitored under real-life driving conditions; however, FOTs require significant time and resources and do not allow the researchers to test the limits of the technology. In order to bridge these gaps, the vehicle simulation method has been developed for the evaluation of in-vehicle safety devices. In this method, a vehicle model is initially built and a safety technology model is then established and superimposed onto the vehicle model. Effective modeling of the safety technologies is often difficult because the crash detection algorithms are typically proprietary information. Consequently, hybrid simulation systems which are called hardware-in-the-loop (HiL) have been proposed to integrate actual hardware with the simulation method. The underlying concept of HiL systems is to test actual hardware in a vehicle simulation software. In 1994, Sailer and Essers published a paper that describes the development of a HiL system built to test an antilock brake system (ABS) for trucks (5). In addition, a HiL system was built by the University of Michigan Transportation Research Institute (UMTRI) to test an electronic stability program (ESP) for tractor-semitrailers (6). A HiL system was developed to evaluate the safety benefits of stability control systems for tractor-semitrailers in a previous study (7). In that study, the Bendix electronic stability control (ESC) system was integrated and tested using a HiL system. In relation to the previous study, this study focuses on the evaluation of an in-vehicle safety device utilizing an independently developed HiL system. Specifically, the objectives of this study are: (1) to evaluate the safety benefits of ESC for tractor-semitrailers using a HiL system; (2) to validate and verify the results of a previous HiL study conducted by UMTRI (6) given that both studies were conducted using independently developed HiL systems; and (3) to provide our lessonslearned to researchers and practitioners on how such HiL systems can be developed and used to test various in-vehicle systems. As addressed later, the benefit estimation procedures utilized in this study are adopted from the UMTRI study to provide a comparable comparison between both studies.
Park, Donoughe and Rakha 4 HARDWARE-IN-THE-LOOP SIMULATION SYSTEM Hardware and Software Integration The HiL simulation system used in the study is illustrated in FIGURE 1. For the system construction, a pneumatic tractor-trailer braking system was integrated into the system with a truck simulation engine, TruckSim. The operations of all the other parts of the truck, aside from the braking system, are simulated in real time using TruckSim. The braking system was built based on a stock 2006 Volvo VNL pulling a 2007 Utility MX6000 van trailer, and an additional Bendix Electronic Stability Program (ESP) was integrated with the system to evaluate the performance of stability control systems. The HiL simulation system has two personal computers (PCs) and a National Instruments PXI-8108 in the computer cluster, as illustrated in FIGURE 1. The first PC is used as the HiL Client that allows the researcher to generate simulation scenarios using TruckSim, and monitor simulation runs in real time from the HiL Client software. From this software, the real-time brake pressures for each axle, the treadle pressures, air tank pressures, steering angle, yaw rate, lateral acceleration, torque, engine revolutions per minute, and speed of the vehicle are displayed in real time. The second PC is the HiL Server that connects the HiL Client and the HiL Simulator. The server software checks its queue if there is a request for a simulation run and then transfers the file to the HiL simulator. The PXI-8108 was used for the HiL simulator platform that reads the pressure transducers throughout the simulator hardware and provides the results to the TruckSim simulation dynamic link library (DLL). In summary, once the request is made from the HiL Client, the HiL Server transfers the file. The HiL Simulator resets the ignition line, initializes the TruckSim simulation, steps through the simulation at the specified time step, terminates the TruckSim simulation upon the predefined criteria, and finally creates the output file and sends it back to the HiL Client through the HiL Server. As mentioned earlier, TruckSim was used to model, analyze, and simulate the dynamic behavior of tractor-semitrailers. TruckSim Offline which consists of the TruckSim database, animator, and plotter was used by the researchers to design the model and create simulation runs on the HiL Client PC. TruckSim Real-Time (RT), which is the TruckSim simulation engine DLL, was installed on the HiL Simulator to provide math models in real time. The HiL Simulator software was developed in the National Instruments LabVIEW 8.6.1 environment and was installed on the HiL Simulator for the communication between TruckSim Real-Time and the braking system. The HiL Client software was also developed in the LabVIEW environment for the real-time monitoring of simulation runs. Additional details of the system construction are available in the literature (7). System Validation The HiL system was first tested to ensure that the ESC reacted in a manner consistent with the system specifications. This was done by evaluating the brake pressures exerted by the ESC when a truck traverses a curve at a speed which activates the system. In the maneuver that activates the ESC, it is expected that the ESC will apply a higher modulated brake pressure to the side of the truck that is on the outside of the turn. For the test, a simulation run in which a 199-degree steering maneuver was involved was tested on the HiL system. Since the 199 degree steering maneuver had been used by the Vehicle Research Test Center (VRTC) for their field experiments, the same maneuver was tested on our system so that the simulation results could be validated against these field measurements. The steering input of the 199-degree steering
Park, Donoughe and Rakha 5 maneuver over time is illustrated in FIGURE 2 (a) and (b). Given that the truck makes a turn to the left, the ESC must apply brakes on the right-hand side of the truck to provide the stabilizing moment. FIGURE 2 subplots (c) and (d) present the plots of braking pressure on the first wheel over time as the truck completes the 199-degree steering maneuver (subplots (a) and (b)) and first achieves greater than 6 degrees of roll. Since the UMTRI study defined a rollover as the case that the roll of a truck is greater than 6 degrees, the same criterion was used to compare results across the two studies. The subplots (a) and (c) correspond to the steering input and brake pressure profiles when the truck was traveling at 49.1 km/hr (30.5 mph) with the ESC disabled. The subplots (b) and (d) show the profiles when the truck was traveling at 57.6 km/hr (36.0 mph) with the ESC enabled. FIGURE 2 clearly demonstrates that the ESC initiated a higher brake pressure on the right side of the truck as compared to the left side when it was activated. In addition to the first verification, the TruckSim RT simulation results were validated using truck experimental data provided by the National Highway Traffic Safety Administration (NHTSA). The trucks were tested at VRTC in East Liberty, Ohio and consisted of a Volvo tractor and a Fruehauf van trailer. A TruckSim model that had already been validated with experimental data was provided (8, 9). The Virginia Tech Transportation Institute (VTTI) and VRTC both simulated one truck with two different loading configurations. The different loading configurations had trailer centers of gravity (CGs) located at 2.0 m and 2.3 m. Each loading configuration was placed on the truck and set to attempt a ramp steering maneuver (RSM). The ESC-equipped truck accelerated from a stop until it reached a speed slightly higher than a desired target speed and then coasted down to the target speed. The RSM was triggered once the truck reached the desired target speed. The steering wheel was turned to the left at a rate of 175 degrees per second until it reached 199 degrees. The steering angle was held for 5 s and then was straightened at the same rate at which it began the maneuver, as illustrated in FIGURE 2. The simulation ran at 200 Hz and the data were collected at 20 Hz (every 0.05 s). Each configuration was simulated to determine the speed at which the following events occurred: 1. Highest speed with no wheel lifts 2. Lowest speed with one wheel lift 3. Lowest speed with multiple wheel lifts 4. Lowest speed that creates at least 6 degrees of roll 5. Lowest speed that causes a physical rollover TABLE 1 shows the results of the simulation runs. Given that the results show the increases in the event speeds, the ESC prevents the events from happening until the truck reaches higher speeds. Specifically, the speed that causes the truck to surpass 6 degrees of roll increased by 12 percent when a truck with a 2.0 m CG configuration is equipped with ESC, while it increased by 18 percent for a higher CG configuration truck. Based on the results, the safety benefits seem to be more significant for higher CG configurations. The ESC is analyzed to be able to delay a physical rollover more significantly when compared to the other events because the differences in the event speeds for physical rollover are greater. A physical rollover happened at 64 km/hr (40.0 mph) for a truck with a 2.3 m CG configuration when the ESC was on, which was 27 percent faster than 50.4 km/hr (31.5 mph). The comparison of the speed, yaw rate, and roll angle profiles of the trailer during the RSM clearly demonstrates how the ESC increases the stability of the truck. FIGURE 3 subplots
Park, Donoughe and Rakha 6 (a), (c), (e), and (g) show the profiles when the ESC was off and subplots (b), (d), (f), and (h) show the profiles when the ESC was on. The ESC applies braking on the axles when the ESC system identifies a dangerous situation to decrease the truck s speed instantly, as illustrated in FIGURE 2. As seen in subplots (d), (f), and (h), the ESC reduces the speed, yaw rate, and roll angle back into a safe range with the exception of the physical rollover case. The speeds that VRTC determined for the aforementioned events were compared against the speeds that were produced by the real-time system. TABLE 2 summarizes the results of the simulations and experiments conducted by VTTI and VRTC. The simulation results of the realtime runs were consistent with the results of VRTC simulations and experiments. In addition, the plots of truck speed, lateral acceleration, yaw rate, roll angle, and vertical force acting on the wheel for each real-time simulation run were reviewed by the VRTC experts who conducted the field experiments. They concluded that all the simulated responses were similar to what they found in the field experiments and the HiL simulator was deemed valid for the simulation analysis. SIMULATION DESIGN Truck Modeling The truck that was modeled in TruckSim for this research is based on a Volvo tractor pulling a Fruehauf box trailer. This truck was modeled because field experimental data from NHTSA and VRTC were available to validate this model. As mentioned earlier, the truck was equipped with a braking system from a 2006 Volvo VNL in combination with a 2007 Utility MX6000 van trailer and Bendix Electronic Stability Program. There were two box-shaped loads that were placed on the trailer to modify the truck s CG. The front load weighed 10,464.4 kg and the rear load weighed 8,650 kg. These loads were selected and located above the trailer axles to create realistic values for the pitch, roll, and yaw of the trailer inertia. In order to increase or decrease the CG of the truck the loads were simply raised or lowered without changing their weights. For the experiment, the truck was configured with CGs at 2.0 m and 2.3 m to model high and medium CGs. The braking capacity of the truck model was consistent with S-Cam brakes that have a maximum torque of 7,500 N-m at a maximum brake pressure of 0.8 MPa. Testing Procedures To run the simulations, the truck was set to accelerate from a stopped position until it reached the desired speed for the simulation. This occurred on a straight-line section of the designed roadway before the truck reached the beginning of a curve. During this acceleration period, small sinusoidal steering maneuvers were completed so that the ESC was able to identify and estimate the trailer load based on its own algorithm. Once the truck reached the desired speed, it was held constant at that speed until the truck reached the point in the road where the curve began. At this point, the truck s throttle was dropped down to zero for the remainder of the maneuver. In other words, the truck began coasting at the beginning of either the 68- or the 227-meter radius curve. The reason that the truck was tested on the 68- and 227-meter radius curves is because the rollover cases selected from the Large Truck Crash Causation Study (LTCCS) data were classified into one of two bins: radius less than 100 m, and radius greater than 100 m. The details of the radius selection are available in a report from UMTRI (6). The average radius of the cases with a radius of less than 100 m was 68 m and 227 m was the average radius of the cases in the other bin. FIGURE 4 illustrates a testing procedure on the 68-meter radius curve. In order to
Park, Donoughe and Rakha 7 estimate the critical speed at which the truck s roll angle surpasses 6 degrees of roll in the simulation, the simulations were run at increasing speeds until the critical speed was identified. Note that the driver was set to steer the truck along the curve of the road to keep the truck on the intended path. TruckSim s steering algorithm optimized the steering input and avoided any sudden or unnecessary corrections. Simulation Results In summary, trucks with CGs located at 2.0 m and 2.3 m above ground-level were tested on the 68- and 227-meter radius curves. FIGURE 5 shows the results of one of the iterative simulation runs when the truck with a 2.0-meter CG height enters the 68-meter curve without the ESC enabled. The minimum y coordinate is - 136 m (68 m 2) and the target speed is 61.3 km/hr (38.3 mph). The highest roll angle is 6.03 degrees and occurs around the 33 rd s. Please note that the curve has a spiral curve section so the truck can enter the curve smoothly. That is why the highest roll angle does not occur at the beginning of the curve. All of the iterative simulation runs were done using the HiL system and critical speeds were determined, as shown in TABLE 3. In addition to these results, the critical speeds determined by UMTRI for the ESC system evaluation are also shown in TABLE 3. Under the test conditions given in this report, the truck with a medium CG (2.0 m) showed that it can enter the 68-meter curve 9.8 km/hr (6.1 mph) faster than can the same truck with the Bendix ESC disabled. For the truck with a higher CG (2.3 m), the ESC-equipped truck can enter the 68-meter curve at 5.6 km/hr faster than if it were not equipped with the ESC. For the 227-meter curve testing, the differences in the critical speed were not as significant as they were for the 68-meter curve testing. These results were different from what was observed in the field experiments conducted by VRTC. The differences observed in the experiments were higher when compared to the results of the current study. Further research is needed to explain the reason. The results of the UMTRI study are similar to those of the current study in showing that the trucks with the lower CG locations are more affected by the stability systems as compared to the trucks with a higher CG; however, the magnitude of the difference in critical speeds from the anti-lock braking system (ABS)-only case to the ESC case is greater in the UMTRI model as compared to what was found here. It is worth to note that these differences are not necessarily associated with differences in the WABCO and Bendix systems given that the two HiL systems were independently built by different research teams using different parts and software. If the performance of the two systems needs to be compared fairly, the two systems should be tested on the same HiL system. However, the results do indicate that possible savings in such systems given that both HiL systems are totally independent. BENEFIT ESTIMATION As mentioned earlier, UMTRI conducted a similar study that quantified the safety benefits of an ESC using their HiL system integrated with the WABCO ESC. This study followed the same procedures used by the UMTRI team per NHTSA s recommendation. Additionally, the crash totals and expert panel rates were obtained from the UMTRI study, as illustrated in FIGURE 6. In the diagram, the gray shapes illustrate the data obtained from the UMTRI study and the blue shapes show what has been calculated by this study. The benefit estimation procedures can be divided into two steps. The first step is to derive total truck rollover and loss of control (LOC) crashes that may potentially benefit from the ESC. The second step is to calculate crash prevention ratios when the ESC technologies are used. Crash reductions are computed by multiplying the 5-year average crash rates by the crash
Park, Donoughe and Rakha 8 prevention ratios. Finally, cost benefits are estimated by multiplying the crash reductions by cost per crash values. The following sections briefly describe the overall procedures and some details regarding the computation of crash prevention ratios using the HiL simulation results. The details of the safety benefit estimation procedures are available in the literature (6, 10). As illustrated in the diagram, total rollover and LOC crashes and injuries were first calculated from the General Estimates System (GES) and TIFA databases and then updated based on the review of the LTCCS database. Specifically, the total rollover and LOC crashes and injuries were categorized by roadway alignment, surface condition, and type of injuries. The crash prevention ratio was calculated in parallel. A total of 159 rollover and LOC crash cases were selected from the LTCCS database to represent all the crash types identified from the GES and TIFA databases. The selected LTCCS cases provided supporting data for the HiL simulation and engineering judgments. Given the LTCCS cases, two rating approaches were used to estimate the effectiveness of the ESC: an expert panel and a HiL simulation. Twenty-two cases were selected for the HiL simulation to estimate the effectiveness of the ESC. The expert panel rated the effectiveness of the the remaining 137 cases based on the proposed rating method because they were found to be too complex to simulate due to the lack of information supporting simulation. To estimate the effectiveness of ESC, this study used the newly constructed HiL simulation system for the 22 cases as conducted by UMTRI. For the other 137 cases, this study used the effectiveness measures given by the UMTRI scientists to provide a direct comparison. Effectiveness Estimated from the HiL Simulation This study used a data set of 3460 lateral acceleration measurements to compute the effectiveness based on the HiL simulation results. The lateral acceleration measurements were taken under normal driving conditions for a Roll Stability Adviser (RSA) FOT study (6, 11). FIGURE 7 (a) shows a histogram of the lateral acceleration measurements and a fitted normal distribution that is superimposed to the measurements. Since the lateral acceleration measurements were taken during normal driving, rollover and LOC crashes did not occur at this lateral acceleration range. Thus, the distribution was shifted to the right, the side of higher lateral acceleration levels, based on a distribution of rollovers on curves from the LTCCS data to include the ranges in which rollover and LOC crashes occurred. Specifically, 46.5 percent of all trucks that rolled over entered the curve (radius < 100 m) below the critical speed based on the LTCCS data, while 75 percent of the trucks entered the curve (radius > 100 m) below the critical speed. Consequently, for the 68-meter curve radius, the distribution was shifted so that the critical speed became the 46.5 th percentile in the shifted distribution. For the 227-meter curve radius the distribution was shifted so that the critical speed became the 75 th percentile. Given the shifted distribution, the area of the lateral acceleration range in which the ESC prevents rollovers was calculated under the assumption that the lateral acceleration data are normally distributed. The effectiveness was then calculated by dividing the area by the probability of rollover and LOC crashes when the ESC was not used. For the evaluation of ESC, the lateral acceleration distribution and procedures used by UMTRI were employed. First, the critical speeds obtained from the HiL simulation were converted to lateral acceleration levels, as shown in TABLE 3. To compute lateral acceleration for a given vehicle speed the following equation was used.
Park, Donoughe and Rakha 9 Ay 2 v 3.6 = (1) gr Where, v is the vehicle speed in km/hr, g is the gravitational constant (9.81 m/s 2 ), r is the curve radius in meters, and Ay is the lateral acceleration in g. Given the lateral accelerations, the lateral acceleration distribution was shifted following the previously described procedures. For instance, the critical speed for the 68-meter radius curve is 56.8 km/hr at the CG of 2.3 m when the ESC is not activated, and that is equivalent to a lateral acceleration of 0.37 g. Hence, the lateral acceleration distribution is shifted to the right so that 0.37 g becomes the 46.5 th percentile, as illustrated in FIGURE 7 (b). The effectiveness of the ESC is then calculated by dividing the area A by the summation of the areas A and B since the summation of the areas A and B is the ratio of speeds at which trucks not using the ESC rollover and the area B is the ratio of speeds at which trucks using the ESC rollover. The details of the area and effectiveness are shown in TABLE 4. The effectiveness of the ESC on the 22 LTCCS cases was determined based on the curve radius and height of the CG. Cost-Benefit Analysis Given the effectiveness measures for the 159 LTCCS cases, the mean effectiveness measures were calculated by road alignment and surface condition for the rollover and LOC crashes. The mean effectiveness measures were then multiplied by the total number of relevant rollover and LOC crashes that were previously calculated. In order to estimate the monetary safety benefits, the gross benefits of Bendix ESC, the numbers of reduced crashes were multiplied by the unit costs of crashes that were obtained from the literature (12). The total reduction in crashes, deaths, and injuries is equivalent to $1.663 billion in 2007 dollars. Given that the total cost-benefit estimated by UMTRI is $1.738 billion in 2007 dollars, it is 4.4 percent less than the benefit that UMTRI reported for the WABCO ESC. In terms of the number of crashes prevented, VTTI estimated 4.1 percent less than the UMTRI estimate. Given that it is reported that 2,617,118 combination trucks were registered in 2009 in the US, the gross benefit per truck will be $635/year (13). If the lifetime of a semi-trailer truck is assumed to be 10 years then the benefit is $6,354 over the lifetime of the truck. FINDINGS AND CONCLUSIONS In conclusion, ESC on tractor-semitrailers is expected to provide significant safety benefits. The estimated reductions are 4462 crashes, 121 deaths, and 5669 injuries and these reductions are equivalent to $1.663 billion in crash-related costs. When compared to the UMTRI study results, the number of reduced crashes is 4.1 percent less and the annual economic benefit is 4.4 percent less. However, these results do not imply that the WABCO ESC system outperforms the Bendix ESC system since there are various factors that affect the HiL simulation results. For example, one of the factors would be that the two different teams used different versions of TruckSim and different hardware setups for the simulation. Therefore, it is not recommended that the results of the current study be used to compare the two systems; instead the results can be a verification of the previous reported results. While the method of using a HiL system provides an innovative approach to determining the benefits of truck rollover systems, it is a very complex setup. Each signal from the simulator to the real-time machine must be in exactly the right format and timing to ensure that all of the inputs are being accounted for in the simulation run. Since most safety systems do not release
Park, Donoughe and Rakha 10 their detection algorithms, creating the correct signal is often an iterative task. Many verification tests need to be run with field data before the results of a HiL simulator test are deemed reliable. It was found that variations in weather, such as cold to warm temperatures, can sometimes have an effect on the performance of the hardware in the loop. It is advised that researchers using a HiL system perform a standard simulation run to verify that the simulator is calibrated correctly each time the system is used. Once the HiL system is properly tested and validated, then it can be a reliable and convenient method to evaluate in-vehicle safety devices without comparing the results with test-track testing results. The system can also be extended to evaluate other invehicle systems without the need to conduct field tests. REFERENCES 1. NHTSA, Traffic Safety Facts: Large Trucks. DOT HS 811 158 [Brochure], 2008. 2. Wang, J. and F.M. Council, Estimating truck-rollover crashes on ramps by using a multistate database. Transportation Research Record, 1999. 1686(Compendex): p. 29-35. 3. Baker, D., R. Bushman, and C. Berthelot, Effectiveness of truck rollover warning systems. Transportation Research Record, 2001. 1779: p. 134-140. 4. Winkler, C.B. and R.D. Ervin. On-board estimation of the rollover threshold of tractor semitrailers. in 16th IAVSD Symposium on the Dynamics of Vehicles on Roads and Tracks. 1999. Pretoria, South Africa. 5. Sailer, U. and U. Essers, Real-time simulation of trucks for hardware-in-the-loop applications. SAE Technical Paper 942297, 1994. 6. Woodrooffe, J., D. Blower, T. Gordon, P.E. Green, B. Liu, and P. Sweatman, Safety Benefits of Stability Control Systems for Tractor-Semitrailers. October, 2009, DOT HS 811 205, National Highway Traffic Safety Administration, Washington, D.C.,. p. 160. 7. Donoughe, K., H. Rakha, W. Swanson, S. Park, and J. Bryson, Development of a Hardware-in-the-Loop Testbed for Evaluating Truck Safety Systems. Presented at 90th Annual Meeting of the Transportation Research Board, Washington D.C., 2011. 8. Mcnaull, P., D. Guenther, G. Heydinger, P. Grygier, and M.K. Salaani, Validation and Enhancement of a Heavy Truck Simulation Model with an Electronic Stability Control Model. SAE Technical Paper 2010-01-0104, 2010. 9. Chandrasekharan, S., D. Guenther, G. Heydinger, M. Salaani, and S. Zagorski, Development of a Roll Stability Control Model for a Tractor Trailer Vehicle. SAE International Journal of Passenger Cars-Mechanical Systems, October 2009. 2(1): p. 670-679. 10. Woodrooffe, J., D. Blower, and P.E. Green, Tractor Trailer Rollover Prevention: The Effectiveness of Electronic Stability Control Systems. Presented at 90th Annual Meeting of the Transportation Research Board, Washington D.C., 2011. 11. Winkler, C., J. Sullivan, S. Bogard, R. Goodsell, and M. Hagan, Field Operational Test of the Freightliner/Meritor WABCO Roll Stability Advisor & Control at Praxair, in UMTRI report UMTRI-2002-24. 12. Zaloshnja, E. and T. Miller, Unit Costs of Medium/Heavy Truck Crashes / Final Report for Federal Motor Carrier Safety Administration. 2006, Pacific Institute for Research and Evaluation,. 13. FHWA, Highway Statistics 2009. 2011.
Park, Donoughe and Rakha 11 LIST OF TABLES TABLE 1 Event Speeds for RSM TABLE 2 Comparison of VRTC and VTTI Values for Validation Purposes TABLE 3 Critical Speeds Determined by the HiL Simulation TABLE 4 Effectiveness Calculation based on the HiL Simulation TABLE 5 Estimated Total Cost-Benefits (in 2007 dollars) LIST OF FIGURES FIGURE 1 Schematic diagram of the Hardware-In-the-Loop system. FIGURE 2 Brake pressure on the first axle over time. FIGURE 3 Steering input with ESC-off (a) and ESC-on (b), speed with ESC-off (c) and ESC-on (d), yaw rate with ESC-off (e) and ESC-on (f), and roll angle with ESC-off (g) and ESC-on (h) of the trailer configured with the 2.3 m center of gravity. FIGURE 4 Illustration of a test on the 68-meter curve. FIGURE 5 Plot of Y coordinate, speed, and roll angle on the 68-meter curve with a center of gravity of 2.3 m. FIGURE 6 Potential safety benefit estimation procedure diagram. FIGURE 7 (a) Histogram of lateral acceleration measured during normal driving and (b) illustration of effectiveness calculation for the testing of the 68-meter curve and CG 2.3 m.
Park, Donoughe and Rakha 12 TABLE 1 Event Speeds for RSM Classification Outcome ESC-off ESC-on 2.0-m Trailer Center of Gravity Height 2.3-m Trailer Center of Gravity Height No Wheel Lift 51.2 km/hr (32.0 mph) 56.8 km/hr (35.5 mph) Single Wheel Lift 52.0 km/hr (32.5 mph) 57.6 km/hr (36.0 mph) Multiple Wheel Lift 52.8 km/hr (33.0 mph) 58.4 km/hr (36.5 mph) Achieves > 6 Roll 53.6 km/hr (33.5 mph) 60.0 km/hr (37.5 mph) Physical Rollover 58.4 km/hr (36.5 mph) 68.8 km/hr (43.0 mph) No Wheel Lift 47.2 km/hr (29.5 mph) 53.6 km/hr (33.5 mph) Single Wheel Lift 48.0 km/hr (30.0 mph) 54.4 km/hr (34.0 mph) Multiple Wheel Lift 48.8 km/hr (30.5 mph) 55.2 km/hr (34.5 mph) Achieves > 6 Roll 48.8 km/hr (30.5 mph) 57.6 km/hr (36.0 mph) Physical Rollover 50.4 km/hr (31.5 mph) 64.0 km/hr (40.0 mph)
Park, Donoughe and Rakha 13 Classification TABLE 2 Comparison of VRTC and VTTI Values for Validation Purposes VTTI TruckSim Real-Time 2.0-m Trailer CG Height with ESC-off VTTI TruckSim Real-Time 2.3-m Trailer CG Height with ESC-off VRTC TruckSim Results from June 2009 VRTC Experimental: 53-ft. Strickland Box Trailer GVWR* Load *GVWR gross vehicle weight rating Highest Speed with No Tip-Up 51.2 km/hr (32.0 mph) 47.2 km/hr (29.5 mph) 49.6 km/hr (31.0 mph) 48.0 km/hr (30.0 mph) Lowest Speed with Tip-Up 52.0 km/hr (32.5 mph) 48.0 km/hr (30.0 mph) 50.4 km/hr (31.5 mph) 49.6 km/hr (31.0 mph)
Park, Donoughe and Rakha 14 TABLE 3 Critical Speeds Determined by the HiL Simulation Classification UMTRI Current Study Curve radius 68 m 227 m Center of Gravity Without ESC (km/hr) With ESC (km/hr) Difference (km/hr) Without ESC (km/hr) With ESC (km/hr) Difference (km/hr) 2.0 m 62.8 80.5 17.7 61.2 71.0 9.8 2.3 m 57.9 70.8 12.9 56.8 62.4 5.6 2.0 m 109.4 120.7 11.3 107 109.7 2.7 2.3 m 101.4 106.2 4.8 99.1 101.4 2.3
Park, Donoughe and Rakha 15 TABLE 4 Effectiveness Calculation based on the HiL Simulation Curve Center of UMTRI Current Study Radius Gravity Effectiveness (%) Area A Areas A & B Effectiveness (%) 68 m 2.0 m 99.9 0.531 0.535 99.2 2.3 m 99.9 0.441 0.535 82.5 227 m 2.0 m 95.8 0.102 0.25 40.7 2.3 m 61.8 0.083 0.25 33.2
Park, Donoughe and Rakha 16 TABLE 5 Estimated Total Cost-Benefits (in 2007 dollars) Crash Type Total Fatal A-injury B-injury C-injury No injury Injury unknown Rollover 1,459,363,952 712,614,007 406,472,769 235,937,180 90,021,355 14,122,629 196,012 LOC 203,151,450 96,841,826 37,046,724 37,656,294 24,029,023 7,525,066 52,517 Sum 1,662,515,402 809,455,833 443,519,493 273,593,474 114,050,378 21,647,695 248,529
Park, Donoughe and Rakha 17 FIGURE 1 Schematic diagram of the Hardware-In-the-Loop system.
Park, Donoughe and Rakha 18 FIGURE 2 Brake pressure on the first axle over time.
Park, Donoughe and Rakha 19 FIGURE 3 Steering input with ESC-off (a) and ESC-on (b), speed with ESC-off (c) and ESC-on (d), yaw rate with ESC-off (e) and ESC-on (f), and roll angle with ESC-off (g) and ESC-on (h) of the trailer configured with the 2.3 m center of gravity.
Park, Donoughe and Rakha 20 50 40 Time (s) 30 Moving Direction 20 Radius = 68 m 10 0-50 Y Coordinate -100-150 -100-50 0 50 X Coordinate 100 150 FIGURE 4 Illustration of a test on the 68-meter curve.
Park, Donoughe and Rakha 21 50 Y Coordinate 0-50 -100 Beginning of Curve -150 0 5 10 15 20 25 30 35 40 45 50 Time (s) 80 Speed (km/hr) 60 40 20 Target Speed= 61.3km/hr 0 0 5 10 15 20 25 30 35 40 45 50 Time (s) 5 Roll (Degree) 0-5 Roll Angle=-6.03-10 0 5 10 15 20 25 30 35 40 45 50 Time (s) FIGURE 5 Plot of Y coordinate, speed, and roll angle on the 68-meter curve with a center of gravity of 2.3 m.
Park, Donoughe and Rakha 22 Portion Obtained from the UMTRI Report Compute crash totals from GES and TIFA databases Select 159 LTCCS Cases Portion Updated by This Study Update the crash totals based on the LTCCS review Expert panel rates the effectiveness of 137 cases 22 cases rated based on HiL simulation results Five-year annual average crashes Crash prevention ratio (%) Compute crash reduction Estimate cost benefit FIGURE 6 Potential safety benefit estimation procedure diagram.
Park, Donoughe and Rakha 23 FIGURE 7 (a) Histogram of lateral acceleration measured during normal driving and (b) illustration of effectiveness calculation for the testing of the 68-meter curve and CG 2.3 m.