Rethinking Energy Management for Autonomous Ground Robots on a Budget

TL;DR

The paper addresses energy-constrained autonomous ground robots by jointly optimizing computing frequency $f$ and locomotion speed $s_{max}$ under a fixed energy budget $E_b$. It introduces the Predictable Energy Consumption Controller (PECC), which uses DNN-based predictors and a Gurobi optimization to select $(f, s_{max})$ from a finite set while enforcing a per-meter energy budget $E_{bpm}$ and a safety-driven reaction distance $RD^{des}$. Two variants, PECC-δ (dynamic updates) and PECC (static), are evaluated against an energy-efficient baseline through real-world trials and Gazebo simulations, showing travel-time gains up to 17% real-world and 31% in simulation, with energy usage close to the budget (about 95% and 91%). The results demonstrate that budget-aware cyber-physical optimization can substantially improve AGR performance within energy constraints, enhancing task completion time and fleet availability without compromising safety.

Abstract

Autonomous Ground Robots (AGRs) face significant challenges due to limited energy reserve, which restricts their overall performance and availability. Prior research has focused separately on energy-efficient approaches and fleet management strategies for task allocation to extend operational time. A fleet-level scheduler, however, assumes a specific energy consumption during task allocation, requiring the AGR to fully utilize the energy for maximum performance, which contrasts with energy-efficient practices. This paper addresses this gap by investigating the combined impact of computing frequency and locomotion speed on energy consumption and performance. We analyze these variables through experiments on our prototype AGR, laying the foundation for an integrated approach that optimizes cyber-physical resources within the constraints of a specified energy budget. To tackle this challenge, we introduce PECC (Predictable Energy Consumption Controller), a framework designed to optimize computing frequency and locomotion speed to maximize performance while ensuring the system operates within the specified energy budget. We conducted extensive experiments with PECC using a real AGR and in simulations, comparing it to an energy-efficient baseline. Our results show that the AGR travels up to 17\% faster than the baseline in real-world tests and up to 31\% faster in simulations, while consuming 95\% and 91\% of the given energy budget, respectively. These results prove that PECC can effectively enhance AGR performance in scenarios where prioritizing the energy budget outweighs the need for energy efficiency.

Figures (3)

• Figure 1: Experimental setup used in our experiments: (a) real-world map, (b) map for simulations, (c) our prototype AGR, and (d) example obstacles in the real map.
• Figure 2: Motivation results: (a) Total energy consumed by the AGR while navigating the paths AB and BA on the map shown in Figure \ref{['fig:real_map']}, (b) average power consumption, (c) end-to-end execution time, and (d) maximum reaction distance at different computing frequencies and maximum achievable speed.
• Figure 3: Comparison of relative travel time and energy budget usage between EE and PECC across different reaction distances, energy budgets and paths for (a) real-world experiments using the map shown in Figure \ref{['fig:real_map']}, and (b) simulations using the map shown in Figure \ref{['fig:sim_map']}.