20 July 2020 in IEEE Robotics and Automation Letters
This paper deals with the problem of multiple robots working together to explore and gather at the global maximum of the unknown field. Given noisy sensor measurements obtained at the location of robots with no prior knowledge about the environmental map, Gaussian process regression can be an efficient solution to construct a map that represents spatial information with confidence intervals. However, because the conventional Gaussian process algorithm operates in a centralized manner, it is difficult to process information coming from multiple distributed sensors in real-time. In this work, we propose a multi-robot exploration algorithm that deals with the following challenges: 1) distributed environmental map construction using networked sensing platforms; 2) online learning using successive measurements suitable for a multi-robot team; 3) multi-agent coordination to discover the highest peak of an unknown environmental field with collision avoidance. We demonstrate the effectiveness of our algorithm via simulation and a topographic survey experiment with multiple UAVs.
11 June 2020 in IEEE Control Systems Letters
This letter presents a combination of reinforcement learning (RL) and deterministic controllers to learn a quadrotor control. Learning the quadrotor flight in a standard RL approach requires many iterations of trial and error, which may bring about risky exploration and battery consumption. In this letter, we integrate a classical controller such as PD (proportional and derivative) or LQR (linear quadratic regulator) with a RL policy using their linear combination. The proposed method is not only simple to use, but also fast in learning convergence. When the algorithm is evaluated for a quadrotor trajectory tracking by means of a velocity control for both simulation and experiment, it demonstrates the faster convergence rate and better control performance in comparison with an existing rapid model-based RL method.
17 January 2020 in IEEE Robotics and Automation Letters
This paper presents real-time optimization algorithms on trajectory generation and control for a multi-rotor with a suspended load. Since the load is suspended through a cable without any actuator, movement of the load must be controlled via maneuvers of the multi-rotor, which brings about difficulties in operating this system. Additionally, the highly nonlinear dynamics of the system exacerbates the difficulties. While trajectory generation and control are essential for safety, energy efficiency, and stability, the aforementioned characteristics of the system add challenges. With this in mind, the authors propose real-time path planning and optimal control algorithms for collision-free trajectory generation and trajectory tracking. For the dynamics, simplified dynamic models of the system are proposed by considering time delay in attitude control of the multi-rotor. For collision avoidance, the vehicle, cable, and load are considered as ellipsoids with different sizes and shapes, and collision-free constraints are expressed in an efficient and nonconservative way. The augmented Lagrangian method is applied to solve the nonlinear optimization problem with the nonlinear constraints in real-time. For control of the system, model predictive control with a sequential linear quadratic solver is used. Several simulations and experiments are conducted to validate the proposed algorithm.
29 August 2022 in IEEE Transactions on Vehicular Technology
In this paper, we propose a distributed beam tracking scheme for millimeter-wave unmanned aerial vehicle (UAV)-assisted communication systems. In the considered systems, aerial node base (aNB) connects the ground vehicles (GV) through directive beams to mitigate the path loss. However, line-of-sight (LoS) blockage may result in the link disconnection, which leads to the corruption of observations. To solve this issue, we extend the one-to-one beam tracking into the multi-node cooperative beam tracking scheme. For the multi-node cooperative problem, we develop the distributed extended Kalman filter (EKF)-based beam tracking to overcome the high nonlinearity and field-of-view (FoV) limitation of the observation model. While the multi-node cooperation scheme alleviates the corruptions of the observations, it may cause the error propagation between neighbor nodes. Thus, we incorporate the resilient network control into the proposed scheme. Consequently, a distributed EKF-based multi-node cooperative beam tracking protocol is proposed for the correction of corrupted observations and link stability. From our comprehensive simulation, the result of the proposed scheme shows 87.07% error reduction compared to the single node method. This improvement is comparable to the result of the centralized method (91.28% error reduction), and the proposed scheme also has the advantage of scalability.
29 August 2023 in International Journal of Control, Automation, and Systems
1 March 1995 in Sensors and Actuators B: Chemical