Robert Bosch Australia: Advice on automated and zero emission vehicle infrastructure for Infrastructure Victoria Executive Summary The Bosch Group is a leading global supplier of technology and services and one of the world s largest suppliers to the global automotive industry. Bosch believes the future of mobility will be electrified, automated and connected. Highly automated driving is considered an integral part of the future of mobility and Bosch is significantly invested in vehicle systems development. Infrastructure requirements are an integral part of future adoption of highly automated and electrified vehicles. The Bosch experience proposes an iterative approach to determining infrastructure requirements based on the development, projected availability and market release of vehicles over the next ten to fifteen years. Bosch expects aspects of highly automated driving to be available in vehicles by 2025. An important part will be access to appropriate testing and certification facilities for highly automated driving vehicles. There are yet no international standards that encompass a whole of life approach to highly automated driving however the Bosch experience outlined below highlights use cases that may be considered important in defining infrastructure requirements. Bosch as a mobility solutions provider The advent of highly automated driving is being driven by Bosch and similar organisations. Consequently the Bosch Group is making decisions with far reaching strategic consequences, some of which are leading to the largest new investments in the company s history, others to significant restructuring. In 2017, Bosch and Daimler demonstrated an automated valet parking solution in Stuttgart where a Mercedes E-Class vehicle is able to park itself from entrance of the multi-story carpark without human intervention. Bosch and Mercedes also jointly announced they will bring robo taxis to Stuttgart in 2022. Bosch has also announced a 300 million investment in artificial intelligence that will work on extending the capability of artificial intelligence to automated driving vehicles. The Bosch Group has more patents than any other company in the fields of automated driving. Bosch in Australia In Australia, some 200 Bosch engineers work on global automotive projects. The Bosch vehicle safety division is at the forefront of automated driving having locally developed and built Page 1 of 6
Australia s most advanced highly automated driving vehicle (SAE Level 3). The demonstration vehicle is a partnership with the Victorian Government, TAC and VicRoads and is being used in development trials and demonstrations by government agencies and private companies to inform the development of regulations and infrastructure. The operation of highly automated vehicles on Victorian Roads Incremental development and uptake of Advanced Driver Assistance Systems (ADAS) and Automated Driving (AD) technology is well underway. While changes to travel behaviours and implications to transport networks remain speculative, Bosch believes that multimodal, connected transport and ride-sharing trends will proliferate quickly as the intelligent transport megatrend continues to mature. Potential benefits and risks are well documented and Bosch monitors these closely as it relates to safety and commercial positioning. Highly automated driving will develop evolutionarily (not revolutionary). As a result, infrastructure requirements will need to be responsive and agile until highly automated driving becomes the norm. ADAS technology, the precursor to highly automated driving, is broad and already exists in new vehicles. 2017 European new vehicles have between 25-50% ADAS technology which is expected to reach 100% by 2025. As a result, Bosch expects highly automated driving to be available by 2025. Page 2 of 6
The evolution of ADAS systems will result in a mixed fleet of automated and non-automated vehicles on roads. The interface between these vehicles will affect infrastructure requirements in the short-term especially as vehicle manufacturers compete for market release. Bosch Australia s demonstration vehicle has been granted exemption to be used on Victorian roads. As a result, we have gathered significant data on how the vehicle interacts with its environment that can be shared with appropriate authorities and regulatory agencies. It is clear to us that a verification process is required in the early stages of highly automated driving that takes into account the current diverse range of road infrastructure. Bosch supports the need for testing in private use environments, both simulated and real world. Tests that need to be conducted to inform infrastructure requirements include the following: Intersections with signals Variety of curve radii and road gradients Communications infrastructure for connectivity testing Cross walks, lane delineators, curb cuts, bike lanes, grade crossings Various road surfaces (concrete, asphalt, gravel) Two and three lane road Variety of signage and traffic control devices Hydrants, footpaths, etc. Simulated weather environments Round-about and tunnels Fixed, variable street lighting Pedestrians, bicycles and buildings (fixed and moving) Page 3 of 6
Key learnings from Bosch s existing trials have determined the following: Communications from external data sources to the vehicle alone will not be sufficient in safety critical incidents; the on-board sensor suite should remain the primary input Compatibility and standardization across topics such as software, communication frequencies, nomenclature, signage, etc. should be considered and prioritized Assuming that the aim of AD technology is for it to function as safely, reliably and efficiently as possible, the introduction of any uncertainties and irregularities to the operating system greatly diminishes the chance of success Other challenging infrastructure scenarios Bosch has encountered are outlined in the following table. Category Topics Notes Lane markings Signage Visibility Materials Colour Visibility Frame rates Variable signs Detection rates Test with different configurations Materials used Activation conditions Detection of ghost objects, false positives Barriers Centre /side barriers Gutter height Materials used Localization vs Perception Road topology Other road infrastructure Curvature Camber Surfaces Incline / decline Moving objects Temporary features Motion control- transfers trajectory to actuation Localization based on road features Free space detection Environment modelling Tunnels & Bridges Roadworks Localization testing Robustness Perception Page 4 of 6
Category Topics Notes Nomenclature Highway conditions Classification / semantic segmentation Objects / Roads Deep learning Behaviour planning & prediction Validation of corner cases Tunnels Connectivity Satellite and mobile signals Lighting and visibility V2V / V2X FOTA Remote diagnostics Cyber / data security As a result, Bosch believes further verification and validation of vehicle behaviour is a key input into infrastructure requirements for the adoption of highly automated driving. A purpose-built facility is currently being considered by Bosch and industry partners to accelerate the learnings and understanding of impacts to road infrastructure. Ownership and market models of highly automated vehicles Bosch recognises the movement towards shared and multi-modal transport. Ownership and market models of highly automated vehicles will foreseeably be different to traditional vehicle ownership models. Bosch has invested in new mobility platforms for vehicle usage (i.e. e!system, an electric vehicle range and data source platform) and user platforms (i.e. MyScotty, an aggregator of shared mobility transport encompassing ride-share, scooter/bike share and public transport). While electric vehicles may require the installation of public charging, automated driving vehicles in a shared mobility system may reduce the burden on infrastructure with vehicles potentially reducing the burden on parking spots (i.e. return to base or automated, close knit carparks). Zero emissions vehicles Electrification is a key pillar of the future of mobility for Bosch. Although traditional combustion powertrains are evolving through a phase of alternative, carbon reduced fuels or hybrids, the evolution towards pure electric technologies will become commercially sustainable as the Page 5 of 6
technology matures. Every global vehicle manufacturer is developing pure electric or partially electric models. In regards to infrastructure, a typical passenger electric vehicle will require 7-8kWh of energy per day to meet transport needs. This energy is inherently flexible and if managed well can improve the efficiency of the energy grid. Current vehicle charging development trends are designed to receive partial, interrupted or as required vehicle energy. Vehicle energy storage will increase resulting in less frequent charging and more flexibility for drivers to schedule charging at the more cost-sensitive or efficient times. For most vehicles, home charging may be adequate. For high power or time critical charging, cloud based vehicle and infrastructure connectivity solutions will play a role. Dynamic forecasting and energy grid demand can manage driver s expectations and requirements for suitable charging options. Therefore, zero emissions vehicle infrastructure require robust connectivity either cloud based or between vehicles and charging systems (outside home plug-in). With local public site network connectivity, energy requirements of aggregations of electric vehicles can be managed within site infrastructure limits. With broader connectivity, aggregations of electric vehicles can act as both loads and temporary sources of energy able to dynamically respond to grid limitations or as a de-centralised power grid. Conclusion Bosch believes the future of mobility will be electrified, automated and connected. New shared mobility or multi-modal transport systems have emerged and will continue to disrupt traditional vehicle ownership. In order to adequately assess infrastructure requirements of highly automated vehicles, an iterative verification and validation process beyond traditional vehicle compliance is required. The advent of highly automated driving is an evolutionary process through the adoption of ADAS technology in vehicles and continual testing of these vehicles as they are developed. RBAU/OFE C/CCR-AU Page 6 of 6