研究路径跟踪线性规划支持向量机(path following linear programming support vector machine,PF-LPSVM)分类算法,利用路径跟踪法求解线性规划的高效性,提高线性规划支持向量机在大规模数据集上的学习效率。给出线性规划支持向量机的...研究路径跟踪线性规划支持向量机(path following linear programming support vector machine,PF-LPSVM)分类算法,利用路径跟踪法求解线性规划的高效性,提高线性规划支持向量机在大规模数据集上的学习效率。给出线性规划支持向量机的模型并将其标准化,导出用路径跟踪法求解线性规划向量机的关键公式,给出完整的算法流程。在随机数据集及UCI数据集上,将所提算法与LibSVM和牛顿法线性规划向量机(Newton-LPSVM,N-LPSVM)做比较,实验结果表明,所提算法用路径跟踪法提高LPSVM的学习效率是可行的,其适用于大规模数据集的学习。展开更多
To resolve the path tracking problem of autonomous ground vehicles,an analysis of existing path tracking methods was carried out and an important conclusion was got.The vehicle-road model is crucial for path following...To resolve the path tracking problem of autonomous ground vehicles,an analysis of existing path tracking methods was carried out and an important conclusion was got.The vehicle-road model is crucial for path following.Based on the conclusion,a new vehicle-road model named "ribbon model" was constructed with consideration of road width and vehicle geometry structure.A new vehicle-road evaluation algorithm was proposed based on this model,and a new path tracking controller including a steering controller and a speed controller was designed.The difficulties of preview distance selection and parameters tuning with speed in the pure following controller are avoided in this controller.To verify the performance of the novel method,simulation and real vehicle experiments were carried out.Experimental results show that the path tracking controller can keep the vehicle in the road running as fast as possible,so it can adjust the control strategy,such as safety,amenity,and rapidity criteria autonomously according to the road situation.This is important for the controller to adapt to different kinds of environments,and can improve the performance of autonomous ground vehicles significantly.展开更多
文摘研究路径跟踪线性规划支持向量机(path following linear programming support vector machine,PF-LPSVM)分类算法,利用路径跟踪法求解线性规划的高效性,提高线性规划支持向量机在大规模数据集上的学习效率。给出线性规划支持向量机的模型并将其标准化,导出用路径跟踪法求解线性规划向量机的关键公式,给出完整的算法流程。在随机数据集及UCI数据集上,将所提算法与LibSVM和牛顿法线性规划向量机(Newton-LPSVM,N-LPSVM)做比较,实验结果表明,所提算法用路径跟踪法提高LPSVM的学习效率是可行的,其适用于大规模数据集的学习。
基金Project(90820302)supported by the National Natural Science Foundation of China
文摘To resolve the path tracking problem of autonomous ground vehicles,an analysis of existing path tracking methods was carried out and an important conclusion was got.The vehicle-road model is crucial for path following.Based on the conclusion,a new vehicle-road model named "ribbon model" was constructed with consideration of road width and vehicle geometry structure.A new vehicle-road evaluation algorithm was proposed based on this model,and a new path tracking controller including a steering controller and a speed controller was designed.The difficulties of preview distance selection and parameters tuning with speed in the pure following controller are avoided in this controller.To verify the performance of the novel method,simulation and real vehicle experiments were carried out.Experimental results show that the path tracking controller can keep the vehicle in the road running as fast as possible,so it can adjust the control strategy,such as safety,amenity,and rapidity criteria autonomously according to the road situation.This is important for the controller to adapt to different kinds of environments,and can improve the performance of autonomous ground vehicles significantly.
