ABSTRACT
This paper proposes a multipurpose reinforcement learning based low-level multirotor unmanned aerial vehicles control structure constructed using neural networks with model-free training. Other low-level reinforcement learning controllers developed in studies have only been applicable to a model-specific and physical-parameter-specific multirotor, and time-consuming training is required when switching to a different vehicle. We use a 6-degree-of-freedom dynamic model combining acceleration-based control from the policy neural network to overcome these problems. The UAV automatically learns the maneuver by an end-to-end neural network from fusion states to acceleration command. The state estimation is performed using the data from on-board sensors and motion capture. The motion capture system provides spatial position information and a multisensory fusion framework fuses the measurement from the onboard inertia measurement units for compensating the time delay and low update frequency of the capture system. Without requiring expert demonstration, the trained control policy implemented using an improved algorithm can be applied to various multirotors with the output directly mapped to actuators. The algorithm's ability to control multirotors in the hovering and the tracking task is evaluated. Through simulation and actual experiments, we demonstrate the flight control with a quadrotor and hexrotor by using the trained policy. With the same policy, we verify that we can stabilize the quadrotor and hexrotor in the air under random initial states.
Subject(s)
Algorithms , Neural Networks, Computer , Computer Simulation , LearningABSTRACT
This paper presents the prism-type holographic optical element (PT-HOE) design for a small-form-factor (SFF) optical pickup head (OPH). The surface of the PT-HOE was simulated by three steps of optimization and generated by binary optics. Its grating pattern was fabricated on the inclined plane of a microprism by using the standard photolithography and specific dicing procedures. The optical characteristics of the device were verified. Based on the virtual image method, the SFF-OPH with the device was assembled and realized.
ABSTRACT
Technological obstacles to the use of rotary-type swing arm actuators to actuate optical pickup modules in small-form-factor (SFF) disk drives stem from a hinge's skewed actuation, subsequently inducing off-axis aberrations and deteriorating optical quality. This work describes a dual-stage seesaw-swivel actuator for optical pickup actuation. A triple-layered bimorph bender made of piezoelectric materials (PZTs) is connected to the suspension of the pickup head, while the tunable vibration absorber (TVA) unit is mounted on the seesaw swing arm to offer a balanced force to reduce vibrations in a focusing direction. Both PZT and TVA are designed to satisfy stable focusing operation operational requirements and compensate for the tilt angle or deformation of a disc. Finally, simulation results verify the performance of the dual-stage seesaw-swivel actuator, along with experimental procedures and parametric design optimization confirming the effectiveness of the proposed system.
