Font Size: a A A

Research On Key Technologies Of Brain Control For Mobile Robot

Posted on:2019-10-24Degree:DoctorType:Dissertation
Country:ChinaCandidate:J S TangFull Text:PDF
GTID:1368330611993070Subject:Control Science and Engineering
Abstract/Summary:PDF Full Text Request
The brain-computer interface(BCI)is a technology that directly translates human thought to the machine commands.Its initial purpose is to provide disabled people with movement assistance to improve the quality of their life.With the development of this technology,the BCI-controlled robotic-arm system and wheelchair system have been proposed to provide upper-limb and lower-limb movement assistance for disabled people.However,these systems still have shortcomings in the following fields: the practicability,the human-machine interaction,the adaptability in some complicated circumstances.In this dissertation,a BCI-based mobile robot architecture is proposed to solve the above problems.On the base of the above mobile robot architecture,theories and applications about the Hybrid BCI,the Asynchronous BCI,the Human-Machine Sharing Analysis,and the Human-Machine Cooperation Design are researched.The work of this dissertation is shown below:1.The software and hardware architecture of the brain-actuated mobile robot are optimized.With integrating an omnidirectional chassis-based wheelchair,a light-weight robotic arm,a Kinect camera,several ultrasonic sensors,and two single-line lidars,etc,the mobile robot plays a role as a robot platform with environmental perception capability.With regard to the software structure,all of the hardware functional modules are implemented with ROS platform to make them work reliably and efficiently.Besides,aiming at decreasing researchers programming burden as well as providing them a light-weight and efficient programming development environment,the BCI-ROS programming framework is designed.This framework has several advantages: The first one is a good faulttolerant and fault-correction capability in EEG signal transmission.The second one is signal event-sign accurate synchronization and optimized GUI programming framework.These works form the basis of the work of this dissertation.2.The optimization design of the P300-BCI,which includes two specific areas:asynchronous P300-BCI and fast-responding P300-BCI,is researched.For the asynchronous P300-BCI designing,the asynchronous-BCI's practical information transfer rate(APITR)algorithm is proposed,which clarifies the optimizing goal for the asynchronous BCI,and further the multi-parameter optimization method to optimize the asynchronous P300-BCI is found.As a scientific and standardized method for designing asynchronous P300-BCI,the performance of the classifier using the above method exceeded its counterparts using the traditional algorithm with approximately 20 percent at most in experiments.With regard to the fast P300-BCI's designing,a new BCI fixed-sequence coding strategy(fixed-sequence + random sequence)was introduced in the P300 paradigm,which is used to resolve the disadvantages,such as slow command updating rate and large command latency,of the traditional oddball P300 paradigm.On the base of the above coding strategy,a new paradigm using siding window signal catching strategy and multiple-scale classifier is introduced,with its command updating rate being 0.2 seconds.This new paradigm provides user immediate feedback and shortens the latency of the system.The simulated BCI experiment based on this paradigm demonstrated its feasibility and effectiveness in real system's control.3.By using asynchronous,hybrid BCI technique and the shared-control method,etc,we have implemented two BCI-actuated robot system with a good passing ability and smart,comprehensive serving ability respectively.On the base of the mobile robot architecture,a hybrid brain-computer interface,an environment perception module,and an auto-navigation module are designed to construct a complete and powerful braincontrolled mobile robot driving system.With this system,users are able to independently drive the robot to navigate indoors and outdoors,upstairs and downstairs.It is a great breakthrough compared with the existing research.With the hybrid brain-computer interface,the control law is carefully designed so that the sMI-BCI and P300-BCI alternatively activates,which guarantees a good user experience,and a stable and reliable cooperation between different BCI systems.The research goes deeper because an intelligent braincontrolled mobile robot system is built.In this system,the deep-learning frameworkbased target recognition algorithms and other computer vision algorithms are used to recognize,localize and track the targets of the environment.On the basis of the above work,a decision-making assistant module is designed to smartly help users efficiently output their intent.With regard to the task execution,the auto-navigation,which does not rely on any prior knowledge,is based on the on-site captured grid map.Combined with image tracking technique,the system can work in a non-cooperative(unfamiliar)environment and has certain ability to operate dynamic moving targets.Three patients recruited from rehabilitation hospital and four healthy participants tested this system and all of them completed the designated tasks.The results proved that the system responds quickly in various environments and works reliably and stably.The system has been proved to be a comprehensive,intelligent and friendly brain-controlled robot system,as it is able to provide users with comprehensive life assistance such as mobile cruise and target items operation.To sum up,this dissertation focuses on the implementation of brain-controlled mobile robot systems,and carries out systematic research range from theory,method to technology.The designed smart brain-controlled mobile robot is efficient,intelligent and userfriendly,therefore it is a powerful movement assistance robot system.
Keywords/Search Tags:brain computer interface, P300, brain-actuated rehabilitation veichle, brain-actuated wheelchair, brain-actuated robotic arm
PDF Full Text Request
Related items