A5103022362- Guidance and Control Systems
- Evacuation and Crowd Dynamics
- Robotic Path Planning Algorithms
- Medical Imaging Techniques and Applications
- Advanced MRI Techniques and Applications
- Extremum Seeking Control Systems
- Atomic and Subatomic Physics Research
Northwestern Polytechnical University
2023-2025
In this paper, we adopt the fuzzy actor–critic learning (FACL) and model predictive control (MPC) algorithms to solve pursuit–evasion game (PEG) of quadrotors. FACL is used for perception, decision-making, predicting trajectories agents, while MPC utilized address flight target optimization Specifically, based on information opponent, agent obtains its own strategy by using algorithm. Based reference input from algorithm, algorithm develop altitude, translation, attitude controllers...
Abstract In this paper, we study the multi‐pursuer single‐evader pursuit‐evasion (MSPE) differential game in a continuous environment with consideration of obstacles. We propose novel algorithm based on reinforcement learning and transfer learning. source task stage, employ Q‐learning value function approximation method to overcome challenges posed by large‐scale storage space required conventional Q‐table solution method. This approach expands discrete effectively reduces demand for space....
This paper explores a pursuit–evasion game (PEG) based on quadrotors by combining fuzzy Q-learning (FQL) and model-predictive control (MPC) algorithms. Initially, the FQL algorithm is employed to perceive, make decisions, predict trajectory of evader. Based position velocity information both players in game, pursuer quadrotor determines its action strategy using algorithm. Subsequently, state feedback controller designed MPC algorithm, with reference inputs derived from Within each cycle,...
This paper employs the fuzzy actor-critic learning (FACL) and Kalman filter (KF) to tackle pursuit-evasion game (PEG) within a continuous environment, considering scenario involving multiple pursuers single evader. We design reasonable reward functions for pursuer evader, enabling them complete task achieve obstacle avoidance. The strategies both evader are acquired through FACL algorithm, while is extended from discrete domain domain. Additionally, use KF predict evader's position,...
Purpose: High temporal resolution volumetric thoracic MRI is needed to accurately describe the lung and tumor motion. The main aim of present study substantially increase 3D acquisition speed with k‐space under‐sampling along phase encoding directions subsequent image reconstruction using Compressed Sensing (CS). Method Materials: CS was simulated fully sampled breath‐ holding a cancer patient. imaging 1.8×1.8×2.6 mm. Biased random imposed in by selecting more points near center, fewer...