Action-Aware Pro-Active Safe Exploration for Mobile Robot Mapping

TL;DR

The paper tackles safe exploration and mapping of unknown environments by introducing an action-aware proactive exploration framework that tightly integrates perception, viewpoint selection, and proactive replanning. It defines safe and informative exploration through maximal clearance paths, action-aware frontier viewpoints, and preventive replanning based on immediately available actionable information, combined with verifiably safe unicycle path-following. The approach is validated through numerical simulations and real-robot experiments, showing that geodesic navigation cost and last-mile preventive planning yield the best trade-off between adaptability, safety, and exploration efficiency. The results suggest that navigation cost can dominate information content in dense sensing scenarios and outline directions for extending to multi-viewpoint planning and SLAM under action-aware strategies.

Abstract

Safe autonomous exploration of unknown environments is an essential skill for mobile robots to effectively and adaptively perform environmental mapping for diverse critical tasks. Due to its simplicity, most existing exploration methods rely on the standard frontier-based exploration strategy, which directs a robot to the boundary between the known safe and the unknown unexplored spaces to acquire new information about the environment. This typically follows a recurrent persistent planning strategy, first selecting an informative frontier viewpoint, then moving the robot toward the selected viewpoint until reaching it, and repeating these steps until termination. However, exploration with persistent planning may lack adaptivity to continuously updated maps, whereas highly adaptive exploration with online planning often suffers from high computational costs and potential issues with livelocks. In this paper, as an alternative to less-adaptive persistent planning and costly online planning, we introduce a new proactive preventive replanning strategy for effective exploration using the immediately available actionable information at a viewpoint to avoid redundant, uninformative last-mile exploration motion. We also use the actionable information of a viewpoint as a systematic termination criterion for exploration. To close the gap between perception and action, we perform safe and informative path planning that minimizes the risk of collision with detected obstacles and the distance to unexplored regions, and we apply action-aware viewpoint selection with maximal information utility per total navigation cost. We demonstrate the effectiveness of our action-aware proactive exploration method in numerical simulations and hardware experiments.

Figures (10)

• Figure 1: Action-aware proactive exploration incrementally maps an unknown environment by using safe and informative paths toward the best frontier viewpoints maximizing information utility per total navigation cost, along with systematic preventive replanning based on the immediately available actionable information of viewpoints to avoid redundant last-mile motion. (top) The elements of the action-aware proactive exploration framework. (bottom) An example of safe autonomous robotic exploration with the resulting occupancy map and the robot's exploration trajectory.
• Figure 2: (left) The safe planning space, $\mathcal{F}_{\!\mathrm{plan}}(t)$ (yellow), and control space, $\mathcal{F}_{\!\mathrm{ctrl}}(t)$ (blue), for exploration are constructed by eroding the free space, $\mathcal{F}(t)$ (white), based on the robot's body radius $\rho$ and safety clearance $\varepsilon$. (right) The connected frontier regions (colored dashed lines) are the boundary between the known free space, $\mathcal{F}(t)$ (white), and the unknown space, $\mathcal{U}(t)$ (gray), while the detected obstacle surfaces correspond to the known occupied space, $\mathcal{O}(t)$ (black).
• Figure 3: The local visit cost (left) for safe and informative path planning in exploration is determined as the ratio of the distance to unknown regions (middle) to the distance to collision (right), balancing obstacle clearance while biasing toward unknown regions. In low-cost regions (blue), the robot remains near unexplored areas at a safe distance from collisions.
• Figure 4: A unicycle mobile robot navigates to the best reachable viewpoint of a frontier region, based on the maximum information utility per navigation cost, by safely and persistently following the reference path (red line) using unicycle feedback motion prediction (yellow cone) towards the local control goal (green dot) on the path. (left) The viewpoint sets (colored patches) and the selected viewpoints (colored circles) of the frontier region (colored dashed line). (right) The immediately available actionable information of a viewpoint is determined by its visible frontiers (highlighted in bold).
• Figure 5: Initial robot poses for safe exploration and mapping in numerical simulations using (left) 12m$\times$16m and (right) 28m$\times$19m office-like environments, where the robot's sensing regions are depicted as colored patches.

Theorems & Definitions (7)

• Remark 1
• Remark 2
• Lemma 1
• proof
• Remark 3
• Proposition 1
• proof