Zijian Wang

Neural controller for swimming modes and gait transition on an ostraciiform fish robot

This paper proposes a novel control strategy with the central pattern generator (CPG) model for motion control of an ostraciiform fish robot. Compared with previous methods, this method can generate diverse swimming modes and manage gait transition with only two inputs. The CPG enables the robot to achieve agile swimming in the three-dimensional space and to automatically switch between the pectoral and caudal gaits with different speeds. Immediate application of the CPG model is illustrated on a boxfish robot propelled by two pectoral fins and one caudal fin, each of which is individually driven by a motor. With a vision sensor integrated in the CPG-based feedback control, the robot performs a target tracking behavior to further validate stability and availability of the model in real time. Both numerical simulation and laboratory experimental results are provided to validate the effectiveness of the proposed CPG model.

Multi-robot Manipulation Without Communication

This paper presents a novel multi-robot manipulation algorithm which allows a large number of small robots to move a comparatively large object along a desired trajectory to a goal location. The algorithm does not require an explicit communication network among the robots. Instead, the robots coordinate their actions through sensing the motion of the object itself. It is proven that this implicit information is sufficient to synchronize the forces applied by the robots. A leader robot then steers the forces of the synchronized group to manipulate the object through the desired trajectory to the goal. The paper presents algorithms that are proven to control both translational and rotational motion of the object. Simulations demonstrate the approach for two scenarios with 20 robots transporting a rectangular plank and 1000 robots transporting a piano.

Multi-robot manipulation with no communication using only local measurements

This paper presents a novel approach to coordinate the manipulation forces of a group of robots without explicit communication during a cooperative manipulation task. Robots use the measurements of the motion of the object as the only information to reach a consensus on their forces. It is proven that the consensus can be reached even if all the robots have different velocity and acceleration measurements since they take measurements at different attachment points around the object while the object is rotating and translating. The convergence of the leader-following process where a leader robot actively steers the forces of all follower robots to navigate the object along a desired trajectory is also proven with Lyapunov stability arguments. We verify our method in both numerical simulations and a physics simulator, where we transport a grand piano with 1001 robots.

Force-Amplifying N-robot Transport System Force-ANTS for cooperative planar manipulation without communication

We propose the concept of a Force-Amplifying N-robot Transport System (Force-ANTS) to coordinate the manipulation forces from a group of robots in order to transport a heavy object in a planar environment. Our approach requires no explicit communication among robots. Instead, we prove that robots can use local measurements of the object’s motion at their attachment points as implicit information for force coordination. A leader (either a robot or human) can guide the whole group towards the destination by applying a relatively small force, whose effect is amplified by the follower robots as they align their forces with the leader’s. Two Force-ANTS implementations are introduced and analyzed, accounting for two different classes of object dynamics: small objects where kinetic friction dominates, and large objects where inertia and viscous friction dominate. Our approach can be used as a modular system for transporting heavy objects of various sizes in many real-life applications. Simulations with up to 1000 robots and experiments using four custom-built robots are conducted to validate our approach. We also conduct human–robot cooperation experiments where the human force is amplified by three follower robots.

Fast, On-line Collision Avoidance for Dynamic Vehicles Using Buffered Voronoi Cells

This letter presents a distributed collision avoidance algorithm for multiple dynamic vehicles moving in arbitrary dimensions. In our algorithm, each robot continually computes its buffered Voronoi cell (BVC) and plans its path within the BVC in a receding horizon fashion. We prove that our algorithm guarantees collision avoidance for robots with single integrator dynamics. We show that our algorithm has computational complexity of

Multirobot Object Transport via Robust Caging

O(k)

OuijaBots: Omnidirectional Robots for Cooperative Object Transport with Rotation Control Using No Communication

, which is the same as that of the optimal reciprocal collision avoidance (ORCA) algorithm, and is considerably faster than model predictive control (MPC) and sequential convex programming (SCP) based approaches. Moreover, ORCA and MPC-SCP require relative position, velocity, and even other information, to be exchanged over a communication network among the robots. Our algorithm only requires the sensed relative position, and therefore is well suited for on-line implementation as it does not require a communication network, and it works well with noisy relative position sensors. Furthermore, we provide an extension of our algorithm to robots with higher-order dynamics like quadrotors. We demonstrate the capabilities of our algorithm by comparing it to ORCA in multiple benchmark simulation scenarios, and we present results of over 70 experimental trials using five quadrotors in a motion capture environment.

Intercepting Rogue Robots: An Algorithm for Capturing Multiple Evaders With Multiple Pursuers

In this paper, we propose a control algorithm to collectively transport an object using a group of relatively low-cost robots. We address this problem using the robust caging, which features reliable object closure with minimum number of robots, and requires no high-precision control capability on the individual robot. Given a 2-D convex object, the proposed method uses the quality of complete robustness to first optimize the number of robots in the initial formation, and then reorient and move the formation. The method is free of force analysis, and therefore less prone to sensor errors and failures. Compared with state-of-the-art multirobot object transport approaches, which require more robots and rely heavily on high-precision control, such as force and torque feedback control, our method uses fewer robots and has high tolerance to control noises. We performed both simulation and real-time experiments to demonstrate the performance of our method. We conclude that the proposed robust caging is promising under reduced number of robots and a certain level of control noises in multirobot object transport tasks.

Kinematic multi-robot manipulation with no communication using force feedback

We propose a distributed force and torque controller for a group of robots to collectively transport objects with both translation and rotation control. No explicit communication among robots is required. This work goes beyond previous works by including rotation control and experimental demonstrations on a custom built robot platform. We prove that follower robots can synchronize both their forces and torques to a leader (either a robot or human) that guides the group, and thus contribute positively to the transport. We introduce a custom-designed omnidirectional robot platform, called the OuijaBot, with sensing and actuation capabilities for cooperative manipulation. Our approach is verified by experiments with four OuijaBots successfully transporting and rotating a payload through a narrow corridor.

Pipelined Consensus for Global State Estimation in Multi-Agent Systems

We propose a distributed algorithm for the cooperative pursuit of multiple evaders using multiple pursuers in a bounded convex environment. The algorithm is suitable for intercepting rogue drones in protected airspace, among other applications. The pursuers do not know the evaders’ policy, but by using a global “area-minimization” strategy based on a Voronoi tessellation of the environment, we guarantee the capture of all evaders in finite time. We present a decentralized version of this policy applicable in two-dimensional (2-D) and 3-D environments, and show in multiple simulations that it outperforms other decentralized multipursuer heuristics. Experiments with both autonomous and human-controlled robots were conducted to demonstrate the practicality of the approach. Specifically, human-controlled evaders are not able to avoid capture with the algorithm.