Integrated Design Framework for Perception System of Automated Electric Vehicles with Enhanced Sensor Reusability

This paper presents an integrated design framework for perception system of automated electric vehicles, which can efficiently generate a comprehensive design scheme for perception system while satisfying the detection demands of multiple automated driving features. Sensor reusability is achieved in the framework, facilitating reduction in system cost and improvement in hardware utilization efficiency. To achieve this, a general feature-oriented distributed design approach based on swarm intelligence optimization is established to generate multiple optional design schemes for each feature, with varied trade-offs on performance and cost. Design efficiency is promoted by this distributed design mechanism. To incorporate the distributed design into a system-level scheme, a fuzzy reasoning control approach is proposed to analytically reason and select the most suitable scheme in each feature by balancing the performance and cost, where schemes containing reusable sensor configurations are selected with priority. By actively guiding the fuzzy reasoning process with the delicately designed control, the dynamic reasoning system is asymptotically stabilized at the expected design scheme. System design scheme is generated by integrating all selected feature-oriented schemes. Afterwards, a system-level redundancy elimination is conduct to further enhance sensor reusability. The effectiveness of the proposed framework is validated through a simulation case study.