Effective deployment of multi-robot teams requires solving several interdependent problems at varying levels of abstraction. Specifically, heterogeneous multi-robot systems must answer four important questions: what (task planning), how (motion planning), who (task allocation), and when (scheduling). Although there are rich bodies of work dedicated to various combinations of these questions, a fully integrated treatment of all four questions lies beyond the scope of the current literature, which lacks even a formal description of the complete problem. In this article, we address this absence, first by formalizing this class of multi-robot problems under the banner Simultaneous Task Allocation and Planning with Spatiotemporal Constraints (STAP-STC), and then by proposing a solution that we call Graphically Recursive Simultaneous Task Allocation, Planning, and Scheduling (GRSTAPS). GRSTAPS interleaves task planning, task allocation, scheduling, and motion planning, performing a multi-layer search while effectively sharing information among system modules. In addition to providing a unified solution to STAP-STC problems, GRSTAPS includes individual innovations both in task planning and task allocation. At the task planning level, our interleaved approach allows the planner to abstract away which agents will perform a task using an approach that we refer to as agent-agnostic planning. At the task allocation level, we contribute a search-based algorithm that can simultaneously satisfy planning constraints and task requirements while optimizing the associated schedule. We demonstrate the efficacy of GRSTAPS using detailed ablative and comparative experiments in a simulated emergency-response domain. Results of these experiments conclusively demonstrate that GRSTAPS outperforms both ablative baselines and state-of-the-art temporal planners in terms of computation time, solution quality, and problem coverage.
27 December 2016 in IEEE Transactions on Automation Science and Engineering
An algorithm called adaptive-neural-intention estimator (ANIE) is presented to infer the intent of a human operator's arm movements based on the observations from a 3-D camera sensor (Microsoft Kinect). Intentions are modeled as the goal locations of reaching motions in 3-D space. Human arm's nonlinear motion dynamics are modeled using an unknown nonlinear function with intentions represented as parameters. The unknown model is learned by using a neural network. Based on the learned model, an approximate expectation-maximization algorithm is developed to infer human intentions. Furthermore, an identifier-based online model learning algorithm is developed to adapt to the variations in the arm motion dynamics, the motion trajectory, the goal locations, and the initial conditions of different human subjects. The results of experiments conducted on data obtained from different users performing a variety of reaching motions are presented. The ANIE algorithm is compared with an unsupervised Gaussian mixture model algorithm and an Euclidean distance-based approach by using Cornell's CAD-120 data set and data collected in the Robotics and Controls Laboratoy at UConn. The ANIE algorithm is compared with the inverse LQR and ATCRF algorithms using a labeling task carried out on the CAD-120 data set.
Complex, multi-task missions require the coordination of heterogeneous robots at multiple inter-connected levels, such as coalition formation, scheduling, and motion planning. This challenge is exacerbated by dynamic changes, such as sensor and actuator failures, communication loss, and unexpected delays. We introduce Dynamic Iterative Task Allocation Graph Search (D-ITAGS) to simultaneously address coalition formation, scheduling, and motion planning in dynamic settings involving heterogeneous teams. D-ITAGS achieves resilience via two key characteristics: i) interleaved execution, and ii) targeted repair. Interleaved execution enables an effective search for solutions at each layer while avoiding incompatibility with other layers. Targeted repair identifies and repairs parts of the existing solution impacted by a given disruption, while conserving the rest. In addition to algorithmic contributions, we derive accurate bounds on schedule suboptimality and provide insights into the inherent trade-off between time and resource optimality in these settings. Our experiments reveal that i) D-ITAGS is significantly faster than recomputation from scratch in dynamic settings, with little to no loss in solution quality, and ii) the theoretical bounds on optimality gap consistently hold in practice.
19 March 2018 in International Journal of Intelligent Robotics and Applications