Sensor Fusion-Based Autonomous Indoor Mobile Robot Navigation Using Fuzzy-Based CEKF

Autonomous mobile robots have gained significant traction in various fields, such as industrial automation, logistics and healthcare, where they face complex challenges in dynamic indoor environments. The need for precise localization, while contending with obstacles, changing ambient conditions and real-time decision-making, makes localization an intricate task. This research presents a novel approach for enabling self-navigating indoor mobile robots. Unlike previous methods, this approach utilizes an advanced Constrained Extended Kalman Filter (CEKF) combined with a fuzzy-based adjustment mechanism to improve sensor data fusion. The robot employs a range of sensors, including odometry, Inertial Measurement Units (IMUs), LIDAR and cameras, to accurately interpret its surroundings. The fuzzy-based CEKF dynamically adjusts the weight of sensor data based on the quality and reliability of the individual measurements, allowing the system to adapt to varying environmental conditions. Simulation results highlight the effectiveness of the proposed model, demonstrating a significant reduction in error compared to earlier models, improving the robot’s localization accuracy. Additionally, different filters were tested and the optimal filter with the lowest error was identified, reinforcing the model’s reliability. Simulation results demonstrate a significant reduction in error compared to earlier models, with a Mean Absolute Error (MAE) of 0.095, a Mean Squared Error (MSE) of 0.0 and a Root Mean Square Error (RMSE) of 0.095. Additionally, the results demonstrate that the fuzzy-based CEKF model outperforms conventional methods, including EKF, and shows reduced computational time compared to other filters, with an operation time of 3[Formula: see text]s.