Aixue Ye

A docking control method in narrow space for intelligent wheelchair

A docking control method which is based on vision feedback and applied in narrow space is proposed in this paper. By it the intelligent wheelchair can be docked into a U-shaped bed automatically and reconfigured to a flat stationary bed. First, The U-shaped bed is identified and its relative pose to the wheelchair is localized according to some features of specific symbol set on the bed. Second, the improved line trajectory tracking algorithm is performed to make the wheelchair to dock into the bed by using vision feedback and dead reckoning. Experiment results validate the feasibility and effectiveness of the proposed method.

A vision-based guidance method for autonomous guided vehicles

Guiding technology is one of the key technologies for the automated guided vehicles (AGV). A vision-based guiding system obtains the environment image with a camera, gets the parameters of the road by processing the image and guides the AGV navigations. Although pasting color or reflecting tape on the ground as a guide line can provide road information for an AGV, but since the tape can not provide position information for the AGV, it is necessary to use other devices to get the position information. A new vision-based guiding method which makes use of a novel artificial landmark called MR (Mobile Robot) code, is presented in this paper, and a localization algorithm using the MR code is also given. Since the paper-made MR codes can be pasted on the path easily, and can provides position and navigation information for an AGV, it can be used easily in a production line. Experiments have also been carried out to show the effectiveness and practicality of the proposed method.

A U-shape bed localization method on embedded vision system

In order to dock an embedded intelligent wheelchair into a U-shape bed automatically through visual servo, this paper proposes a real-time U-shape bed localization method on an embedded vision system based on FPGA and DSP. This method locates the U-shape bed through finding its line contours. The task can be done in a parallel way with FPGA does line extraction and DSP does line contour finding. Experiments show that, the speed and precision of the U-shape bed localization method proposed in this paper which based on an embedded vision system can satisfy the needs of the system.

Docking control of intelligent wheelchair based on vision

A docking control method which is based on vision feedback is proposed in this paper for the intelligent wheelchair which is subject to nonholonomic motion constraints and can reconfigure between bed and wheelchair. First transform the coordinates of the line segments which are extracted in the back-view image from image coordinate system to wheelchair coordinate system by using the homography matrix. According to the Euclid features of the beds contour, identify the U-shaped bed and calculate the posture of the intelligent wheelchair in the bed coordinate system. By using the vision feedback and dead reckoning, posture stabilization and line trajectory tracking are performed to make the intelligent wheelchair to dock into the bed portion precisely. Considering the limited input of the control system, this paper also gives a new control law for posture stabilization. Experiments validate the effectiveness and practicality of the proposed method.

Visual servo control of intelligent wheelchair mounted robotic arm

A visual servo control system of intelligent wheelchair mounted robotic arm is presented, which consists of intelligent human-machine interaction (HMI), visual servo controller and motion controller. An adaptive visual feedback controller based 2D image is designed, which ensure that the manipulator reaches a desired position quickly and grasps a target. With the help of human-machine interaction (HMI), the WMRA (wheelchair mounted robotic arms) autonomously tracks a steady target and grasps the target via visual servo controller. The experimental results show the system has good performances.

Robust Dense Visual Odometry with boundary pixel suppression

Pose estimation and 3D environment reconstruction are crucial for autonomous navigation in mobile robotics. Robust dense visual odometry based on a RGB-D sensor uses all pixels to estimate frame-to-frame motion by minimizing the photometric and geometric error. 3D coordinates of each pixel are calculated necessarily with its corresponding depth. However, depths of some pixels near object boundaries from RGB-D sensors are not accurate. The general robust dense visual odometry does not consider depth noise impact for photometric error and geometric error. In this paper, we construct uncertainties of photometric error and geometric error with depth noise and point out depth noise near object boundaries can significantly affect the result of motion estimation. We present a modified robust dense visual odometry with boundary pixel suppression. Publicly available benchmark datasets are employed to evaluate our system, and results showed that our method achieves higher accuracy compared with the state-of-the-art Dense Visual Odometry (DVO).

A method of self-localization of robot based on infrared landmark

Aiming at the problem of robot self-localization based on vision, this paper proposed a method for robots localization based on infrared landmark. First select the suitable infrared source. Second calculate the camera extrinsic parameters using the algorithm based on P3P. Third give the calculation method of robots pose to the landmark based on the camera extrinsic. Experiment results validate the feasibility and effectiveness of the proposed method.

A reconfigurable wheelchair/bed system for the elderly and handicapped

A Reconfigurable Wheelchair/Bed System (RWBS) for the elderly and handicapped is designed. The control system of this reconfigurable wheelchair consists of a main control unit, a motion control unit, a posture adjusting unit, a physiological signal acquisition unit, a sensors processing unit, an image processing unit and an intelligent human-machine interaction (HMI) unit. With the human-machine interaction (HMI) unit, the wheelchair can be used as a common powered wheelchair, adjust its posture so that the user can change his/her body to an ideal position or to a flat stationary bed. One of the distinguished points of this reconfigurable wheelchair/bed system is that it can be docked into the bed frame automatically by just push a button with the help of a visual feedback control unit. The three control modes can be changed via the touch screen. Posture stabilization and trajectory tracking methods for the visual servoing system are exploited and, experiments are also carried out to show the effectiveness.