Failure is an option: Task and Motion Planning with Failing Executions

Abstract

Task and motion planning represents a powerful set of hybrid planning methods that combine reasoning over discrete task domains and continuous motion generation. Traditional reasoning necessitates task domain models and enough information to ground actions to motion planning queries. Gaps in this knowledge often arise from sources like occlusion or imprecise modeling. This work generates task and motion plans that include actions cannot be fully grounded at planning time. During execution, such an action is handled by a provided humandesigned or learned closed-loop behavior. Execution combines offline planned motions and online behaviors till reaching the task goal. Failures of behaviors are fed back as constraints to find new plans. Forty real-robot trials and motivating demonstrations are performed to evaluate the proposed framework and compare against state-of-the-art. Results show faster execution time, less number of actions, and more success in problems where diverse gaps arise. The experiment data is shared for researchers to simulate these settings. The work shows promise in expanding the applicable class of realistic partially grounded problems that robots can address.

Enabling robustness to failure with modular field robots

“…A recent review on robot failure by Liu et al (2023) showed that most of the research studies focus on operational failures decoupled from system failures. Pan et al (2022) looked at failures in task and motion and adapted their approaches based on the actions that failed during task execution. However, the hardware failures are not explicitly captured or triggered.…”

“…On the adaptive modeling perspective, many recent works have focused on Bayesian inference ( Hammond et al, 2019) , adapting the underlying world transition models ( Pan et al, 2022) and learning the task feasibility model to deal with uncertainties during execution ( Noseworthy et al, 2021) . In terms of application, most of the focus has been on systems recovery via resolving programmatic deadlocks in the service robots ( Das et al, 2021) ; Hammond et al, 2019) or, in the real world, mostly on grasping and bin-picking problems.…”