ABSTRACT: Driving on freeways in free-flow conditions is an important application scenario for autonomous vehicles (AV). This task requires a speed control algorithm that ensures the safety and comfortability of AV occupants, especially on freeways with complex alignments. There are two commonly used speed control algorithm: (1) cruise control (CC) which adopt a constant speed close to the design speed of the freeway, and (2) segment-based speed control (SP) which adjust AV speed according to the current segment’s geometric features. Both algorithms fail to consider the complex speed adjustments of human beings on continuously demanding roadway alignments. To address this limitation, speed selection by human drivers along roads with complex alignments was studied to propose an anthropomorphic AV speed control algorithm (AN) which satisfied AV occupants’ expectation. Considering the diversity of drivers’ speed preference, the clustering method was first applied to identify three different driving styles, conservative, aggressive and experienced. In order to demonstrate the developed AN speed-selection algorithm performance in relation to the CC and AP strategies, the lateral and vertical acceleration rates along complex road alignments were obtained. The results indicated that the AN algorithm adjusted the AV behavior to various driving styles along complex road alignments and thus outperform the other two evaluated algorithms in terms of driving comfortability. A comparison of the algorithm performance on four typical combined alignments indicates the proposed anthropomorphic speed control algorithm could follow the speed preferences of experienced drivers on downslope sections without jeopardizing safety.
Zhigui Chen, Xuesong Wang, Qiming Guo, Andrew P. Tarko. Anthropomorphic Automated Driving Speed Control Algorithm on Complex Alignment Freeways. Transportation Research Board 100th Annual Meeting, Washington D.C., USA, 2021. 1.25-29.