This paper presents control and coordination algorithms for groups of vehicles. The focus is on autonomous vehicle networks performing distributed sensing tasks where each vehicle plays the role of a mobile tunable sensor. The paper proposes gradient descent algorithms for a class of utility functions which encode optimal coverage and sensing policies. The resulting closed-loop behavior is adaptive, distributed, asynchronous, and verifiably correct.

Coverage control for mobile sensing networks

A Glimpse at Set Theory. The Real Numbers. The Topology of Cartesian Spaces. Convergence. Continuous Functions. Functions of One Variable. Infinite Series. Differentiation in RP Integration in RP.

The elements of real analysis (2nd edition), by Robert G. Bartle. Pp xv, 480. £10. 1976. SBM 0 471 05464 X (Wiley)

The recognition that minimizing an integral function through variational methods (as in the last chapters) leads to the second-order differential equations of Euler-Lagrange for the minimizing function made it natural for mathematicians of the eighteenth century to ask for an integral quantity whose minimization would result in Newton’s equations of motion. With such a quantity, a new principle through which the universe acts would be obtained. The belief that “something” should be minimized was in fact a long-standing conviction of natural philosophers who felt that God had constructed the universe to operate in the most efficient manner—but how that efficiency was to be assessed was subject to interpretation. However, Fermat (1657) had already invoked such a principle successfully in declaring that light travels through a medium along the path of least time of transit. Indeed, it was by recognizing that the brachistochrone should give the least time of transit for light in an appropriate medium that Johann Bernoulli “proved” that it should be a cycloid in 1697. (See Problem 1.1.) And it was Johann Bernoulli who in 1717 suggested that static equilibrium might be characterized through requiring that the work done by the external forces during a small displacement from equilibrium should vanish. This “principle of virtual work” marked a departure from other minimizing principles in that it incorporated stationarity—even local stationarity—(tacitly) in its formulation. Efforts were made by Leibniz, by Euler, and most notably, by Lagrange to define a principle of least action (kinetic energy), but it was not until the last century that a truly satisfactory principle emerged, namely, Hamilton’s principle of stationary action (c. 1835) which was foreshadowed by Poisson (1809) and polished by Jacobi (1848) and his successors into an enduring landmark of human intellect, one, moreover, which has survived transition to both relativity and quantum mechanics. (See [L], [Fu] and Problems 8.11 8.12.)

Variational Principles in Mechanics

H. J. Kushner and P. G. Dupuis. Springer-Verlag, New York/Heidelberg, May 1992. 439 pp., DM 98. ISBN 3-540-97834-8

Numerical methods for stochastic control problems in continuous time

Games and Decisions. By R. Duncan Luce and Howard Raiffa. Pp. xi, 509. 70s. 1957. (J Wiley & Sons)

This letter presents a new control paradigm applicable to nonlinear systems such as robots subject to chance and covariance assignment constraints which we refer to as hierarchical optimal covariance control. To the best of our knowledge, this is the first study to formulate the hierarchical optimal covariance control problem involving multiple operational tasks. The framework is defined as a multi-stage optimization problem considering multiple hierarchical tasks specified in lexicographic order. Towards this goal, we first approximate the nonlinear dynamic model of a robot into multiple linear stochastic systems by linearizing the model along given trajectories. We then project these stochastic models onto the null-space of the previous task models for efficiently solving lexicographical optimization. In addition, we specify probability functions to account for chance constraints using the Whittacker M function. We formulate the chance constraints as a positive semi-definite matrix constraint and solve the hierarchical optimal covariance control problem using sequential semi-definite programming. We demonstrate that this procedure yields higher accuracy for multiple hierarchical tasks than employing deterministic operational space control models.

Hierarchical Task-Space Optimal Covariance Control With Chance Constraints

This paper presents a trajectory generation method for contact-constrained robotic systems such as manipulators and legged robots. Contact-constrained systems are affected by the interaction forces between the robot and the environment. In turn, these forces determine and constrain state reachability of the robot parts or end-effectors. Our study subdivides the trajectory generation problem and the supporting reachability analysis into tractable subproblems consisting of a sampling problem, a convex optimization problem, and a nonlinear programming problem. Our method leads to significant reduction of computational cost. The proposed approach is validated using a realistic simulated contact-constrained robotic system.

Trajectory Generation for Robotic Systems with Contact Force Constraints

—This paper proposes Distributed Model Predictive Covariance Steering (DMPCS), a novel method for safe multi-robot control under uncertainty. The scope of our approach is to blend covariance steering theory, distributed optimization and model predictive control (MPC) into a single methodology that is safe, scalable and decentralized. Initially, we pose a problem formulation that uses the Wasserstein distance to steer the state distributions of a multi-robot team to desired targets, and probabilistic constraints to ensure safety. We then transform this problem into a ﬁnite-dimensional optimization one by utilizing a disturbance feedback policy parametrization for covariance steering and a tractable approximation of the safety constraints. To solve the latter problem, we derive a decentralized consensus-based algorithm using the Alternating Direction Method of Multipliers (ADMM). This method is then extended to a receding horizon form, which yields the proposed DMPCS algorithm. Simulation experiments on large-scale prob- lems with up to hundreds of robots successfully demonstrate the effectiveness and scalability of DMPCS. Its superior capability in achieving safety is also highlighted through a comparison against a standard stochastic MPC approach. A video with all simulation experiments is available here.

Distributed Model Predictive Covariance Steering

In this paper, we address a minimum-time steering problem for a drone modeled as point mass with bounded acceleration, across a set of desired waypoints in the presence of gravity. We first provide a method to solve for the minimumtime control input that will steer the point mass between two waypoints based on a continuous-time problem formulation which we address by using Pontryagin’s Minimum Principle. Subsequently, we solve for the time-optimal trajectory across the given set of waypoints by discretizing in the time domain and formulating the minimum-time problem as a nonlinear program (NLP). The velocities at each waypoint obtained from solving the NLP in the discretized domain are then used as boundary conditions to extend our two-point solution across those multiple waypoints. We apply this planning methodology to execute a surveying task that minimizes the time taken to completely explore a target area or volume. Numerical simulations and theoretical analyses of this new planning methodology are presented. The results from our approach are also compared to traditional polynomial trajectories like minimum snap planning.

Smooth time optimal trajectory generation for drones

In this paper, we address a pursuit-evasion game involving multiple players by utilizing tools and techniques from reinforcement learning and matrix game theory. In particular, we consider the problem of steering an evader to a goal destination while avoiding capture by multiple pursuers, which is a high-dimensional and computationally intractable problem in general. In our proposed approach, we first formulate the multi-agent pursuit-evasion game as a sequence of discrete matrix games. Next, in order to simplify the solution process, we transform the high-dimensional state space into a low-dimensional manifold and the continuous action space into a feature-based space, which is a discrete abstraction of the original space. Based on these transformed state and action spaces, we subsequently employ min-max Q-learning, to generate the entries of the payoff matrix of the game, and subsequently obtain the optimal action for the evader at each stage. Finally, we present extensive numerical simulations to evaluate the performance of the proposed learning-based evading strategy in terms of the evader's ability to reach the desired target location without being captured, as well as computational efficiency.

Efstathios Bakolas

Papers

Hierarchical Task-Space Optimal Covariance Control With Chance Constraints

Trajectory Generation for Robotic Systems with Contact Force Constraints

Distributed Model Predictive Covariance Steering

Smooth time optimal trajectory generation for drones

Min-Max Q-Learning for Multi-Player Pursuit-Evasion Games.