Optimal Robot Pose Estimation Using Scan Matching by Turning Function

The turning function is a tool in image processing that measures the difference betweentwo polygonal shapes. We propose a localization algorithm for the optimal pose estimation ofautonomous mobile robots using the scan-matching method based on the turning function algorithm.There are several methodologies aimed at moving the robots in the right way and carrying out theirmissions well, which involves the integration of localization and control. In the proposed method, thelocalization problem is implemented in the form of an optimization problem. Afterwards, the turningfunction algorithm and the simplex method are applied to estimate the localization and orientationof the robot. The proposed algorithm first receives the polygons extracted from two sensors’ data andthen allocates a histogram to each sensor scan. This algorithm attempts to maximize the similarityof the two histograms by converting them to a unified coordinate system. In this way, the estimateof the difference between the two situations is calculated. In more detail, the main objective of thisstudy is to provide an algorithm aimed at reducing errors in the localization and orientation of mobilerobots. The simulation results indicate the great performance of this algorithm. Experimental resultson simulated and real datasets show that the proposed algorithms achieve better results in terms ofboth position and orientation metrics.