Simultaneous Localization and Mapping (SLAM) With Kalman Filtering

Robot self-localization and mapping, or as it is termed Simultaneous Localization and Mapping (SLAM), is a common use case in robot functions. As a complex system that integrates analog sensor based data acquisition and processing SLAM has some accuracy limitations based on the sensors accuracy and environmental conditions that may alter or disrupt sensing [1]. The objective of this project is to demonstrate the benefits of Kalman Filtering on processing of the disruptive or noisy data for the goal of robot localization and mapping. In short Kalman Filter takes the mathematical model of the process and the measurements. It predicts the future state acquires and adjusts the measurements, updates prediction of next states based on the success or errors of the prior prediction. Kalman Filtering is used in broad spectrum of applications including robotics, financial, medical and any other field where there is a need for improved accuracy of measurements or noise reduction. In our application the accuracy of the mapping and localization is greatly dependent on the environmental conditions that may affect the accuracy of the sensors, mechanical and electrical parameters of the hardware and the complexity and dynamics of the mathematical model of the system. In the interest of the scope of this project, for efficiency and maximum rewards vs. efforts we will ignore the environmental variables and focus on the parameters of the process and noisy measurement system. The robot that is used for the project is equipped with laser range scanner, compass and motor encoders. The motion model of the robot is based on differential drive with dual motors one on each side. The laser range scanner and the other sensors are independent and, when fused with Kalman filtering algorithm, will dramatically reduce the inaccuracies of the measurements.