1. Features and challenges of self-driving trucks
Heavy trucks mainly operate on relatively closed expressways, which seem simpler than urban road scenes. However, due to the unique physical characteristics, operating environment and commercial operation requirements of heavy trucks, compared with the automatic driving system of passenger cars, the automatic driving system of heavy trucks has lower perception distance and accuracy of the vehicle, system response speed and accuracy of vehicle control. Higher demands are placed on:
• Long braking distances. The braking distance of a heavy truck at a speed of 100km/h is usually more than 100 meters. Compared with the braking distance of a passenger car within 40 meters, the automatic driving system needs to have a longer perception distance and a faster end-to-end responding speed.
• Huge size and weight. The standard dimensions of the outer contour of the heavy truck are 2.55 meters wide, 4 meters high, and 17 meters long, and the mass of the whole vehicle varies by up to 500% (9 tons to 49 tons) under different trailer loads. The non-rigid body between the front and the trailer Linkage and trailer inertia add to the difficulty of autopilot control. In addition, the operating environment of heavy-duty trucks is complex, and they often face various road environments, such as low-attached ground roads, gravel roads, non-standard lanes, etc., posing greater challenges to the planning and control of heavy-duty truck automatic driving.
• The response speed of the remote control system for heavy trucks is slower and the accuracy is relatively lower. Taking the braking system as an example, heavy trucks generally use air braking technology, which requires a longer pressure building process than hydraulic braking of passenger cars. The response time of the braking system of a heavy truck is usually around 400ms, while that of a passenger car can be within 100ms. Overcoming the delay of the controlled system and realizing precise control poses a relatively large challenge to the automatic driving control system.
• The tolerance range of heavy-duty truck vehicle parameters is large, and with the increase of driving mileage, the vehicle parameters will drift in a wider range. Taking the steering wheel angle offset as an example, when a passenger car leaves the factory, it is within 1°, and when a heavy truck leaves the factory, it can reach 7°. Taking the empty travel of the steering wheel as an example, experimental data shows that for a heavy truck that has been in operation for 12 months, the empty travel has deteriorated by 43% compared with the factory, from an average of 8.4° to more than 12°. The above-mentioned steering wheel parameter tolerance problem requires the automatic driving control system to be able to identify changes in parameters in time and make automatic corrections in time.
• Comprehensive balance requirements in trunk logistics scenarios. The mainline logistics scenario puts forward three requirements for heavy truck operation in terms of safety, timeliness and cost. These three requirements form a triangle relationship in which one ebbs and the other grows. For example, if the pursuit of time efficiency means faster average vehicle speed, it will lead to higher fuel consumption and higher requirements for safety. For the entire autonomous driving system, comprehensive consideration is required to give a global optimal solution under the premise of ensuring safety.
2. The necessity of pre-installation for autonomous driving
(1) Requirements of regulations and standards
In China, in order for a smart truck (assisted driving or automatic driving) to drive on the road and carry out commercial operations legally and compliantly, it must pass the product certification of the Ministry of Industry and Information Technology and other departments in accordance with relevant regulations and standards, and obtain the vehicle product announcement qualification and vehicle operation qualification . Therefore, any plan to purchase vehicles for aftermarket modification can only be used in the test field, and cannot be legally put on the road and carried out commercial operation on a large scale.
(2) Complex vehicle system interaction
As the key system of the vehicle, the automatic driving system requires the cooperation of various system designs at the vehicle level, including the chassis system, powertrain, body system, human-computer interaction system, and networking system. Only the forward R&D mode of pre-installation can meet such complex system interaction requirements, while the modification mode of after-installation can only realize basic functions, and cannot meet the system-wide design requirements necessary for safety and performance. Taking the steering system of L3 level automatic driving as an example, the aftermarket modification mode can only use the traditional steering gear, and the design cannot consider the use scenario of the safety officer holding the steering wheel in the automatic driving mode. The problem that the tracking of the output angle is not in place leads to the safety risk of "drawing the dragon".
(3) Reliability and durability requirements
The operating environment of commercial vehicles has very strict requirements on product reliability and durability. The selection of components and the design and development of vehicle system integration need to meet the corresponding stringent standards and be confirmed through rigorous testing and verification. Taking the camera as an example, in the forward development, on the one hand, the parts themselves must pass a series of mechanical experiments and various environmental experiments; Associated components are jointly designed and tested, and pass the vehicle-level test standard test. Aftermarket products can only focus on the functions of the parts themselves, and cannot fully consider the full-scenario requirements of the automatic driving system in the layout design of the vehicle. For example, the design of heat dissipation and ventilation is missing, which leads to fogging of the front glass of the field of vision, resulting in unsatisfactory system availability. Require.
3. Development principles for pre-installed mass production of self-driving trucks
The development of self-driving vehicles generally starts with a modified proof-of-concept vehicle, but the completion of proof-of-concept and road testing is only
a small step on the road to mass production of self-driving vehicles. In order to achieve large-scale front-loading mass production, it is necessary to meet extensive and stringent requirements in terms of design and verification, mass production and maintenance, and commercialization.