Liying Yang

A robust real-time vision based GPS-denied navigation system of UAV

In 2015 International UAV Innovation Grand Prix the competition, the cargo transport task is assumed as: there are 4 buckets placed in four circles on one moving platform. Firstly, the unmanned aerial vehicle (UAV) is required to identify circle targets and the black and white id marker near the circle on one moving platform, then the UAV chosen a target bucket, tracked and transported it to the other moving platform, until all 4 buckets are transported from one moving platform to the other. In order to accomplish the cargo transport task, a method of a real-time vision-based GPS-denied multiple object tracking for UAV is developed. The Pixhawk controller is used to achieve tracking, that the relative distance and velocity between the target and UAV is estimated by the image. Finally, the experimental results proved the effectiveness and robustness of the algorithm.

GART: An environment-guided path planner for robots in crowded environments under kinodynamic constraints

The problem of three-dimensional path planning in obstacle-crowded environments is a challenge (an NP-hard problem), which becomes even more complex when considering environmental uncertainty and system control. Int this paper, we mainly focused on more challenging problem, that is, path planning in obstacle-crowded environments, and we try to find the relation between contact information and obstacle modeling. We proposed a newactive exploring sampling-based algorithm based on rapidly exploring random tree (RRT), namely, guiding attraction–based random tree (GART). GART introduces bidirectional potential field to redistribute each newly sampled state, such that the in-collision samples can be redistributed for extension. Furthermore, dynamic constraints are deployed to establish forward extending region by GART. Thus, GART can ensure kinodynamic reachability as well as smoothness. Theoretical analysis demonstrate that GART is probabilistic complete, and it obtains faster convergence rate because of its r...

Generation of dynamically feasible and collision free trajectory by applying six-order Bezier curve and local optimal reshaping

This paper considers the problem of generating dynamically feasible and collision free trajectory for unmanned aerial vehicles(UAVs) in cluttered environments. General random-based searching algorithms output piecewise linear paths, which cause big discrepancy when used as navigation reference for UAVs with high speed. Meanwhile, the disturbance may also occur to lead the UAVs into danger. In order to obtain agile autonomy without potential dangers, this paper introduces a three-step method to generate feasible reference. In the first step, a six-order Bezier curve, which uses Tuning Rotation to decrease the curvature, is introduced to smooth the output of the path planner. Then a forward simulation is implemented to find the potential dangerous regions. Finally, the path is reshaped by local optimal reshaping planner to eliminate residual dangers. The three steps form a circulation, the reshaped path sent to the first step again to check dynamic feasibility and safety. The method combining Six-order Bezier curve, Tuning Rotation, and local optimal reshaping is proposed by us for the first time, where the Tuning Rotation is able to meet various curvature requirements without violating the previous path, local optimal reshaping obtains both temporal and spatial reshaping with high time efficiency. The method addresses the system dynamics to achieve agile autonomy, which provides the geometry reference as well as the low level control. The effectiveness of the proposed method is demonstrated by the simulations.

Multi-phase homing optimal control for parafoil system

In this paper an optimal control method for parafoil system homing planning is introduced, which includes multi-phase homing arrangement, optimal homing path calculation using genetic algorithm(GA), and Bezier curves based path planning for parafoil terminal guidance to deal with the situation of variable glide ratios. L1 nonlinear algorithm is adopted to make trajectory tracking. Comparison simulations results show that L1 tracking algorithm has more rapid response speed, higher precision and better wind resistance than PD tracking algorithm in homing path planning with different initial positions and with or without wind.

A Real-Time Multiagent Strategy Learning Environment and Experimental Framework

Many problems in the real world can be attributed to the problem of multiagent. The study on the issue of multiagent is of great significance to solve these social problems. This paper reviews the research on multiagent based real-time strategy game environments, and introduces the multiagent learning environment and related resources. We choose a deep learning environment based on the StarCraft game as a research environment for multiagent collaboration and decision-making, and form a research mentality focusing mainly on reinforcement learning. On this basis, we design a verification platform for the related theoretical research results and finally form a set of multiagent research system from the theoretical method to the actual platform verification. Our research system has reference value for multiagent related research.

Transition control of tilt rotor unmanned aerial vehicle based on multi-model adaptive method

Tilt rotor unmanned aerial vehicle (TRUAV) with ability of hovering and high-speed cruise has attached much attention, but its transition control is still a difficult point because of varying dynamics. This paper proposes a multi-model adaptive control (MMAC) method for a quad-TRUAV, and the stability in the transition procedure could be ensured by considering corresponding dynamics. For safe transition, tilt corridor is considered firstly, and actual flight status should locate within it. Then, the MMAC controller is constructed according to mode probabilities, which are calculated by solving a quadratic programming problem based on a set of input- output plant models. Compared with typical gain scheduling control, this method could ensure transition stability more effectively.

Design and implementation of rotor aerial manipulator system

With the development of rotor aerial robot, it has been widely used as a stabilized flying platform to achieve observation, communication relay and so on by equipping corresponding equipment, which are always called passive tasks. However, many applications need rotor aerial robots to possess more active capability, such as grasp, carry and assemblage. Thus, the demand for an aerial manipulation system gradually appears, and becomes a hot research issue in recent years. Aerial manipulator system, usually composed of rotor aerial robot and the manipulator, is a new type of robot system with active operating capability. Compared with the mobile manipulator based on the ground and underwater mobile robot which is now widely applied, the aerial manipulation has lager working range and better flexibility obviously. Unfortunately, the new aerial manipulation system also brings new challenges, such as strong dynamics coupling, the modeling and stabilization problem, etc. which may affect its further application. Thus, in our work, an actual physical aerial manipulator system is designed and implemented for its system and application research. Based on the system, the preliminary modeling and control was researched and also an experiment for grasping and carry application is conducted. The conclusions of these researches will help us to realize and improve the system in the future.

GPU-based heuristic escape for outdoor large scale registration

Heterogeneous robot introduce a higher perception ability than single type robots in outdoor environments. One key problem is to making the 3D environmental model from the cooperated robots in real time, especially in the unstructured environment. Based on our previous work on outdoor environment registration method, in this paper, we introduce a GPU based Enhanced ICP method for large-scale heterogeneous robot registration. First, we combine the GPU-based nearest neighbor search in the traditional ICP framework. Second, we proposed a measurement and estimation model for the local minima problem. Third, we proposed a GPU-based heuristic escape method to generate the escaping transformation in real time. Experiments involving one unmanned aerial vehicle and one unmanned surface vehicle were conducted to verify the proposed technique. The experimental results were compared with those of normal ICP registration algorithms to demonstrate the performance of the proposed method.

Self-Healing Control Framework Against Actuator Fault of Single-Rotor Unmanned Helicopters

Unmanned helicopters (UHs) develop quickly because of their ability to hover and low speed flight. Facing different work conditions, UHs require the ability to safely operate under both external environment constraints, such as obstacles, and their own dynamic limits, especially after faults occurrence. To guarantee the postfault UH system safety and maximum ability, a self‐healing control (SHC) framework is presented in this chapter which is composed of fault detection and diagnosis (FDD), fault‐tolerant control (FTC), trajectory (re‐)planning, and evaluation strategy. More specifically, actuator faults and saturation constraints are considered at the same time. Because of the existence of actuator constraints, usable actuator efficiency would be reduced after actuator fault occurrence. Thus, the performance of the postfault UH system should be evaluated to judge whether the original trajectory and reference is reachable, and the SHC would plan a new trajectory to guarantee the safety of the postfault system under environment constraints. At last, the effectiveness of proposed SHC framework is illustrated by numerical simulations.

Large-scale stabilization control of input-constrained quadrotor

The quadrotor has been the most popular aircraft in the last decade due to its excellent dynamics and continues to attract ever-increasing research interest. Delivering a quadrotor from a large fixed-wing aircraft is a promising application of quadrotors. In such an application, the quadrotor needs to switch from a highly unstable status, featured as large initial states, to a safe and stable flight status. This is the so-called large-scale stability control problem. In such an extreme scenario, the quadrotor is at risk of actuator saturation. This can cause the controller to update incorrectly and lead the quadrotor to spiral and crash. In this article, to safely control the quadrotor in such scenarios, the control input constraint is analyzed. The key states of a quadrotor dynamic model are selected, and a two-dimensional dynamic model is extracted based on a symmetrical body configuration. A generalized point-wise min-norm nonlinear control method is proposed based on the Lyapunov function, and large-scale stability control is hence achieved. An enhanced point-wise, min-norm control is further provided to improve the attitude control performance, with altitude performance degenerating slightly. Simulation results showed that the proposed control methods can stabilize the input-constrained quadrotor and the enhanced method can improve the performance of the quadrotor in critical states.