Linan Zu

Obstacle avoidance of multi mobile robots based on behavior decomposition reinforcement learning

A reinforcement learning method based on behavior decomposition was proposed for obstacle avoidance of multi mobile robots. It decomposed the complicated behaviors into a series of simple sub-behaviors which were learned independently. The learning structures, parameters and reinforcement functions of every behavior are designed. Then, the fusion for learning results of all behaviors was optimized by learning. This learning algorithm could reduce the status space and predigest the design of reinforcement functions so as to improve the learning speed and the veracity of learning results. Finally, this learning method was adopted to realize the self-adaptation action fusion of mobile robots in the task of obstacle avoidance. And its efficiency was validated by simulation results.

A Far Field Sound Source Localization System for Rescue Robot

A rescue robot can not directly find the target which is in other rooms or invisible places by using visual, ultrasonic or infrared ray sensors, because of the smoke or collapsed walls. According to the diffraction property of audio signal, the sound can bypass obstacles. To find the survivors which are often in far field, an audio-based robot localization system is developed to make it feasible to guide the robot run to the target shouting for help. And ideal effects are obtained in experiments.

Microphone array based auditory localization for rescue robot

Because of the smoke or collapsed walls in the buildings while fire or earthquake occurred, a rescue robot can not directly find the target which is in other rooms or invisible places by using visual, ultrasonic or infrared ray sensors. According to the diffraction property of audio signal, the sound can bypass obstacles. By using microphone array, combined with the speech recognition technology, an audio-based robot navigation system is developed to make it feasible to guide the robot find the target shouting for help. And ideal effects are obtained in experiments.

An auditory system of robot for sound source localization based on microphone array

The mammalian used to direct its vision field to a sound source by its auditory system. So an auditory system of robot for sound source localization is proposed in this paper. Through using the four-element microphone array, the time differences of arrival can be achieved and the sound-source target can be localized. The experiment system is introduced. Then the localization algorithm based on the spherical wave model of propagation is developed in this paper, and the calculation error also be discussed. The experiment results show that the algorithm can accurately calculate the azimuth of the sound source and the localization system can effectively lead the robot to the direction of the target.

Movement recognition by electromyography signal for transfemoral prosthesis control

The surface electromyography signal extracted from the stump was applied to recognize the movement, and be translated into on-off signal to fulfill the conversion from standing to walking. According to the different situation of stump, the signals extracted from different muscles were applied to control the prosthesis. Factor analysis was applied to find the relationship between different muscle groups, and to reduce the number of muscles. Hilbert-Huang transform was applied to extract the frequency features, and time and frequency features were applied to recognize the beginning of barycenter transfer movement. The result of this study indicates that this method can recognize the movement with a higher identification rate.

Electromyogram signal analysis and movement recognition based on wavelet packet transform

For recognizing the movement intent of amputee, surface electromyogram signals which can reflect movement intent and can be measured without invasion were applied to identify movement transition. Wavelet packet transform was applied to analyze the electromyogram signal, extract its frequency feature and recognize movement. The result of this study indicates that if the suitable coefficients were selected, the movement transition from standing to sitting and from sitting to standing can be recognized with a higher identification rate, and has a great potential in practical application of artificial lower limb.

Adaptively target tracking method based on double-Kalman filter in existence of outliers

In the actual course of target tracking, the measurement noise is changing at any time because of the impact of external environment, the problem of the sensors system and other reasons. Meanwhile, there are outliers in the measurements because of the clutter environment. However, the measurement noise covariance is changeless in standard Kalman filter, and the impact of outliers on the tracking results can not be eliminated by Kalman filter, so the tracking results are not perfect inevitably. This article presents a new target tracking method based on double-Kalman filter in existence of outliers, which can solve above problems by adaptively adjusting the measurement noise covariance based on the two results of Kalman filter with the different steps. The presented method can also detect the outliers in accord with the property of the innovation and reconstruct the state estimates to keep the orthogonal property of innovation and reduce the influence of outliers. The simulation results prove that the method can restrain the outliers and increase the accuracy of the target tracking system.

Output feedback control for nonlinear system based on Local Linear Model Trees (LOLIMOT)

A nonlinear adaptive output feedback control approach based on LOLIMOT is proposed for a class of nonaffine system in the non-strict feedback form with unknown nonlinearities in this paper. In order to overcome the disadvantages of output feedback control, LOLIMOT which can identify the nonlinear system accurately is applied. During the control process, backstepping approach is utilized to design the adaptive output feedback controller and the internal pseudocontroller is used as one-step predictor. The effectiveness of proposed scheme is shown by simulation on a proper nonlinear system.

The influence of walking speed on muscle activity of thigh and application in prostheses control

Changes in walking speed may eventually result in modifications of the patterning of muscle activity. The surface electromyography (EMG) signals provide easy and noninvasive access to physiological processes that cause the contraction of muscles. The influence of walking speed on muscle and EMG activity of amputees thigh was analyzed, the speed pattern recognition approach based on the combination of surface EMG signal and knee angle was applied to recognize the walking speed of amputee, and the walking speed control strategy of above knee prostheses was proposed. The result of experiment demonstrated that the phasing of muscle activity remained relatively stable over every walking speed despite substantial changes in EMGs amplitude. Adjusting the swing velocity of prosthesis knee can improve the performance of prosthesis and decrease the energy consumption. The outcome of this investigation could promote the future design of neural-controlled artificial legs.

Research on path planning method of multi mobile robot in dynamic environment

A path planning structure based on hierarchical idea was adopted in this paper. In this structure, the global path planning is done firstly by the robot and the local online adjusting is done afterward. This kind of structure can debase the calculation complexity of path planning. Then, aiming at the problem of local motion planning, a conflict-resolution strategy based on forecast was presented. It can decompose the restrictions of obstacle avoidance problem by establishing rules so as to debase the complexity of the problem. The reasoning process of this local path planning method is simple and practical. The experiments show that the robots can avoid the obstacles which are not only static ones but also moving ones accurately and effectively and the local path planning method is suitable to the conflict-resolution of multi mobile robots.