Coordination of Advanced Driver Assistance Systems and Chassis Systems

Most of chassis systems and Advanced Driver Assistance Systems (ADAS) are developed independently. Car manufacturers have even separate departments to develop the two types of systems. Even though each system category is developed for a different objective, from a global vehicle motion control perspective, these systems may influence the same physical variable. If, from the design process, no supervisory strategy is ensured for both categories, the different systems may interact and generate unwanted behaviors. These interactions can be very unpredictable and may impact the global vehicle safety. Moreover, future automated vehicles will require additional innovative systems. This will make the vehicle over-actuated. The global vehicle motion control should take into account this aspect. New requirements for the longitudinal, transverse, and vertical dynamics should be addressed. Car manufacturers tend to develop tuned rule-based strategies for a specific set of integrated systems by studying different use-cases. On one hand, we cannot foresee all the possible conflicted scenarios. On the other hand, the more numerous the systems get, the more unpredictable the interactions become. The definition of use-cases will become harder or even non-scalable. Our research aims to develop an integrated control architecture where ADAS and chassis systems are optimally coordinated. This our work is based on rather a mathematical formalization of system interactions. This enables developing a multi-layered modular control architecture that starts with a high-level robust control to specify the motion of the car, and then distribute the control on the different implemented systems upstream the low-level controllers using an optimization-based control allocation strategy. This latter strategy is made flexible and extensible so if a new system should be implemented within the same car, the control designer does not have to redesign the overall architecture of the motion control. This remains valid for both ADAS and chassis systems control, and for both trajectory and speed control. To test this, we developed high fidelity vehicle models for different vehicles in Simcenter Amesim, and we implemented the overall control architecture in Matlab, then both softwares can be co-simulated. Through several scenarios, co-simulation results showed that the vehicle performance is improved when activating several systems at the same time with an optimal coordination. More severe situations can be handled at high velocities. In addition, the different systems can be made complementary so if one system fails, another completely different system can takeover the maneuver. The global vehicle is then safer without any redundancy of the advanced expensive systems.