2019/01/22 by Sumit Kumar, Kumar, Sumit, Shushman Choudhary +3
Computer Science · Engineering · #Advanced Image and Video Retrieval Techniques #Artificial Intelligence (cs.AI) #FOS: Computer and information sciences #Human Pose and Action Recognition #Robotic Path Planning Algorithms #Robotics (cs.RO) #Robotics and Sensor-Based Localization
paper · pdf · doi:10.48550/arxiv.1901.07646
openalex publication_date 2019/01/22 · openalex created_date 2022/07/30 · openalex updated_date 2026/07/28
For motion planning in high dimensional configuration spaces, a significant\ncomputational bottleneck is collision detection. Our aim is to reduce the\nexpected number of collision checks by creating a belief model of the\nconfiguration space using results from collision tests. We assume the robot's\nconfiguration space to be a continuous ambient space whereby neighbouring\npoints tend to share the same collision state. This enables us to formulate a\nprobabilistic model that assigns to unevaluated configurations a belief\nestimate of being collision-free. We have presented a detailed comparative\nanalysis of various kNN methods and distance metrics used to evaluate C-space\nbelief. We have also proposed a weighting matrix in C-space to improve the\nperformance of kNN methods. Moreover, we have proposed a topological method\nthat exploits the higher order structure of the C-space to generate a belief\nmodel. Our results indicate that our proposed topological method outperforms\nkNN methods by achieving higher model accuracy while being computationally\nefficient.\n