Energy-Aware Mobile Robot Path Planning Using Dynamic Window Approach
TL;DR
This work extends the Dynamic Window Approach by incorporating a heuristic electromechanical energy model into the local trajectory evaluation and shows that incorporating the energy term improves the smoothness of the commanded motion, as reflected by lower translational and angular velocity variation metrics than the non-energy baseline.
Abstract
Efficient path planning is a key requirement for autonomous mobile robots because it directly affects both navigation performance and operating cost. This work extends the Dynamic Window Approach (DWA) by incorporating a heuristic electromechanical energy model into the local trajectory evaluation. By considering the estimated energy in the cost function, the modified planner can balance heading, obstacle clearance, speed, and energy consumption during local path selection. Simulation results on two test maps show that the proposed extension reduces the total energy consumption by 3.6% and 8.6%, respectively, with an average saving of 6.1% compared with the classical DWA, while also reducing the simulated travel time, the number of planning steps, and the overall elapsed time. A sensitivity analysis further shows that incorporating the energy term improves the smoothness of the commanded motion, as reflected by lower translational and angular velocity variation metrics than the non-energy baseline. In particular, the best tested weighting combination reduced the v-variation from 18.60 to 1.16 and the ω-variation from 34.17 to 9.84, while also achieving the lowest total energy consumption. Although the average power increases, the overall energy decreases because the robot completes the motion in a shorter time. The results also indicate that path length alone is not a sufficient indicator of energy efficiency. Overall, the proposed energy-aware DWA improves mission-level efficiency and yields smoother velocity profiles without compromising successful local navigation.