A scan matching simultaneous localization and mapping algorithm based on particle filter

Purpose - The purpose of this paper is to present a scan matching simultaneous localization and mapping (SLAM) algorithm based on particle filter to generate the grid map online. It mainly focuses on reducing the memory consumption and alleviating the loop closure problem. Design/methodology/approach - The proposed method alleviates the loop closure problem by improving the accuracy of the robot's pose. First, two improvements were applied to enhance the accuracy of the hill climbing scan matching. Second, a particle filter was used to maintain the diversity of the robot's pose and then to supply potential seeds to the hill climbing scan matching to ensure that the best match point was the global optimum. The proposed method reduces the memory consumption by maintaining only a single grid map. Findings - Simulation and experimental results have proved that this method can build a consistent map of a complex environment. Meanwhile, it reduced the memory consumption and alleviates the loop closure problem. Originality/value - In this paper, a new SLAM algorithm has been proposed. It can reduce the memory consumption and alleviate the loop closure problem without lowering the accuracy of the generated grid map.