Engineering firm Ricardo has published a white paper—Key Enablers for the Fully Autonomous Vehicle—highlighting the technologies and development processes that are needed to develop commercially feasible self-driving cars that meet consumer expectations while also achieving compliance with likely future transport regulations.
According to the Boston Consulting Group, the projected size of the global autonomous vehicle market in 2025 will be $36 billion for partially autonomous vehicles (levels 1–3) and $6 billion for fully autonomous vehicles (level 4). This includes both passenger and commercial vehicle uses. The realization of fully autonomous vehicles will require further evolution in software, sensors, integration and efficient system testing beyond what is in place for current advanced driver assistance systems.
The future self-driving vehicle value chain will be driven by software feature sets, complex algorithms for sensor fusion and motion controls, low system costs, and high performance hardware.
… winners [within the automotive industry will be created by the increasing value of software and sensors. Some current players may face a shrinking market for their products due to the consolidation of hardware technologies that enable advanced driver assistance systems (ADAS), and autonomous and connected vehicles.—“Key Enablers”
Ricardo suggests that autonomous vehicle systems will evolve from current decentralized systems which use multiple sensors and ECUs to deliver limited ADAS functions, such as rear parking cameras, adaptive cruise, blind spot detection, lane departure alert, emergency braking and emergency assistance.
By 2020, Ricardo expects that vehicle systems will deliver comprehensive safety using partially centralized systems for self-parking, lane- keeping, smart stop, V2V and vehicle-to-infrastructure (V2I) communication, advanced navigation, WiFi hot spots and a range of assistance with informational and monitoring applications.
By 2025, vehicles will operate with fully centralized systems that incorporate sensor fusion to deliver more features with fewer components (e.g., Audi’s zFAS, earlier post). This type of technology is needed to cost-effectively offer features such as valet parking, vehicle-to-pedestrian communication, augmented-reality driving assistance, smart-routing with estimated time of arrival, and seamless integration with internet devices, Ricardo said.
To successfully accomplish this progression, experienced teams will need to be assembled to address the complexities and develop algorithms for sensor fusion and motion controls to ensure the sensor technology, ECUs and actuation systems are all integrated and working to drive the vehicle. As autonomy continues to evolve, the demand for software will drastically increase, while the hardware proliferation will level off due to system integration.—“Key Enablers”
Another enabling technology to the fully autonomous vehicle is virtual testing and evaluation; extensive testing will be required to develop the necessary software to ensure a Level 4 autonomous vehicle responds appropriately under a wide range of scenarios.Specifically, Ricardo suggests, the industry must leverage agent-based modeling (ABM). This is a simulation methodology that puts agents (vehicles, people or infrastructure) with specific behaviors (selfishness, aggression, etc.) that have connections in a defined environment (cities, test tracks or military installations) to understand the emergent behaviors of the agent during a simulation test.
In terms of advancing virtual testing and validation, the white paper describes Ricardo Agent Drive—simulation software based on ABM methodology that is used to create real-world driving scenarios to test complex driving situations for autonomous vehicles in agent-based simulation. This is under development and will be used to test future autonomous vehicle concepts.
The benefits to car manufacturers and their suppliers that this approach aims to deliver are faster product development cycles, reduced costs related to test-vehicle damage and lower risk of harming a vehicle occupant under test conditions.