Real-time energy-efficient path planning for unmanned ground vehicles using mission prior knowledge

Abstract

Unmanned Ground Vehicle (UGV) missions include situations where a UGV has to choose between alternative paths, and are often limited by the available on-board energy. Thus, we propose a dynamic energy-efficient path planning algorithm that integrates mission prior knowledge with real-time sensory information to identify the most energy-efficient path for mission completion. Our proposed approach predicts and updates the distribution of the energy requirement for alternative paths using recursive Bayesian estimation through two stages: (a) exploration-road segments can be explored to reduce their energy prediction uncertainty; (b) exploitation-the most reliable path is selected using the collected information in the exploration stage and then traversed. Our simulation results show that the proposed approach outperforms offline methods, as well as a method that relies on exploitation only to identify the most energy-efficient path.

Off‐road ground robot path energy cost prediction through probabilistic spatial mapping

“…Ground robot power prediction has often relied on longitudinal vehicle models, perhaps due to their simplicity (Rajamani, 2011;Sadrpour et al, 2014). Sadrpour et al (2014) and Sadrpour, Jin, and Ulsoy (2013) used a longitudinal power model to predict mission energy costs for a ground robot, as well as update those predictions in real-time with collected data. Prior knowledge of factors, including terrain friction, slope, and teleoperator aggressiveness, inform the predictions.…”

“…Sadrpour et al (2014) do not assume constant friction and instead update predicted path energy costs recursively using realtime measurements. Sadrpour et al (2014) consider energy cost predictions for a predefined set of paths, whereas we build a spatial map of the terrain with GPR using data collected during robot operation. The map accounts for changes in terrain friction and is used to predict path energy costs.…”

“…The importance of energy has led to recent research into robot energy characterization, modeling, and prediction (Broderick, Tilbury, & Atkins, 2014;Dogru & Marques, 2018;Sadrpour, Jin, & Ulsoy, 2014). For ground robots operating in off-road environments, terrain plays a significant role in energy costs (Ojeda, Borenstein, Witus, & Karlsen, 2006).…”

“…Three-dimensional environments are challenging because power consumption is likely to be different at the same location, depending on the robot's heading. In addition, longitudinal vehicle models, which are commonly used for robot energy prediction on 3D terrain (Sadrpour et al, 2014;Sun & Reif, 2005), often assume that the friction coefficient (or rolling resistance coefficient) is the same on identical terrain, regardless of slope or direction. Such an assumption can result in biases that throw off predictions.…”