论文部分内容阅读
针对智能汽车的公路轨迹规划问题,本文将最优计算量分配(OCBA)的思想引入基于候选轨迹曲线的规划算法OODE,提出新算法OCBA_OODE。OODE通过比较各候选曲线的“粗糙”(存在偏差但计算量小)评价确定最优轨迹曲线。曲线评价随着投入计算量的增加逐渐收敛至准确值,OODE对各曲线平均分配计算量,OCBA_OODE基于曲线评价循环分配计算量进而提高算法效率。OCBA_OODE在求解质量不下降的前提下,规划速度比OODE的快20%。
In order to solve the problem of road planning in smart cars, this paper introduces the idea of optimal calculation distribution (OCBA) into the OODE algorithm based on candidate trajectory curve, and proposes a new algorithm called OCBA_OODE. The OODE determines the optimal trajectory curve by comparing the “rough” (with a small amount of calculation) of each candidate curve. The curve evaluation gradually converges to the exact value as the amount of calculation increases. OODE distributes the calculation averagely for each curve. OCBA_OODE evaluates the cycle distribution calculation based on the curve to improve the algorithm efficiency. OCBA_OODE In solving the quality does not decline under the premise of planning faster than OODE 20%.