Homepage
From ZXLIwiki
Zexiang Li

Other Activities: 
Publications

Research Interests
 Dexterous manipulation
 Multifingered robotic hand
 Motion planning and control
 Geometric methods of robotics
 Parallel mechanisms
 Electronic manufacturin
 Computeraided set up and inspection
 Precision machining
Recent Projects
Completion of a unified control system architecture for multifingered manipulation
 Abstract
 To accommodate the need of a diverse range of manufacturing assembly/manipulation tasks, a multifingered robotic hand is designed with a complex structure involving:
 a number (three or four) of fingers each with a number (three or four ) of joints;
 a number of sensors of different nature providing information on joint position/torque, contact location/force and geometry/location of the object being manipulated;
 a multiprocessor based control unit. A central issue in the study of multifingered robotic hands is the development of a unified (i.e., hand independent) control system architecture capable of transforming highlevel specifications of a task to lowlevel commands for the fingers.
 In early research at the Robot Manipulation Laboratory of HKUST, we developed the HKUST threefingered robotic hand as a research platform for multifingered manipulation. We proposed a preliminary framework for a unified Control System Architecture for Multifingered Manipulation (CoSAM$^2$). The theory and implementation for several function and sensory modules of CoSAM$^2$ have been completed along with successful demonstration of several important manipulation tasks.
 In this proposal, we propose to broaden and enhance the structure of CoSAM$^2$ through the creation of several new modules:
 Grasp analysis module;
 Dexterous manipulation module;
 Object motion estimation module.
 Utility of these modules will be demonstrated by a series of welldesigned manipulation tasks. It is expected that the integration of these new modules with CoSAM$^2i$ will substantially improve the range of tasks the HKUST hand is capable of performing and thus lead to better understanding of multifingered manipulation.
Automation Technology Cooperative Research Center
 Abstract
 We propose to establish a university/industry cooperative research centre to develop automation technologies for electronic manufacturing. The mission of the Centre is:
 to develop critical automation technologies needed for electronic manufacturing and demonstrate the technologies in prototype products.
 to transfer the technology knowhow to local industries,
 to provide indepth training to local technical staff and engineers in the areas of automation technologies.
 In recent years, Hong Kong's electronics industry has been faced with serious challenges. First, the international market demands high product diversity, rapid product introduction, high product quality and low price.
 Second, the electronics industry had to cope constantly with the problems of rising labor and land costs and high turnover of technical staff, problems that have subsequently caused low assembly yield and poor product quality (See Roadmap for Electronics Packaging and Assembly for Hong Kong, produced by the Hong Kong Electronic Industries Association). In response, many electronics firms have become increasingly interested in adopting stateoftheart automation technologies for their manufacturing operations. But this effort is hindered by several major roadblocks including:
 lack of technology knowhow;
 shortage of experienced engineers and technical staff capable of designing and maintaining modern manufacturing operations; and
 lack of local manufacturers of highend equipment.
 By combining the vast technical expertise and other resources of the University and the important product development experience of the industry, the proposed Centre can help overcome these difficulties. First, an Industry Advisory Board (IAB) will be formed with members from all participating companies of the Centre. Engineers from the member companies will be actively involved in deciding new projects. Second, projects will be initiated to develop core technologies for manufacturing automation. These technologies will be demonstrated in the form of semiproduct (or "productlike") prototypes. Member companies will be kept informed on the status of these projects and will be provided with full access to these technologies. Finally, seminars, workshops and demonstrations will be organized on a regular basis to disseminate these technologies quickly to member companies. By using these technologies, which are already in semiproduct form, member companies can develop their full product prototypes at a substantially reduced cost, cycletime and engineering complexity.
Geometric Algorithms for Hybrid Localization and Tolerance Verification
 Abstract
 Manufacturing enterprises today are confronted by an international market which demands high product diversity, rapid product introduction, high part quality, and low price. These demands create a need for manufacturing systems that are more flexible and agile enough to respond quickly to product and demand changes. In response to this need, computing technologies are increasingly being applied to many aspects of manufacturing.
 Over the last five years, research activities in the Machine Intelligence Laboratory of HKUST have been centered on:
 design of flexible openarchitecture machine tool control systems; and
 development of efficient geometric algorithms for localization of workpieces.
 These technologies address important current manufacturing bottlenecks: the presently excessive costs of setup, refixturing and dimensional inspection. Our overall aim is to extend current onmachine probing inspection capabilities to reduce process time, improve machine tool flexibility, and increase part quality.
 In the present proposal, which supplements our on going laboratory work in an important way, we propose to study efficient algorithms for hybrid localization/envelopment problems, i.e., how to align a part model with a workpiece such that all points measured on finished surfaces of the workpiece match closely to corresponding surfaces on the model while all unmachined surfaces lie outside the model to guarantee the presence of material to be machined at a later time. An important application of hybrid localization/envelopment problems is found in setting up for machining of a partly finished workpiece having both finished and unmachined surfaces, e.g., a casting or forging for post machining in such a way that the model to be cut is inside the material envelope. In parallel, we propose to study geometric techniques for formulating tolerancing notions contained in ANSI Y14.5M, and develop efficient tolerance verification algorithms. By integrating these algorithms with an openarchitecture machine tool environment we hope to complete a ComputerAided Setup and Inspection (CASI) system leading to simplified and accelerated machining cycle with improved product quality.
Geometric Analysis and Implementation of Workpiece Localization Algorithms
 Abstract
 Manufacturing enterprises today are confronted by an international market which demands high product diversity, rapid product introduction, high part quality, and low price. These demands create a need for manufacturing systems that are more flexible and agile enough to respond quickly to product and demand changes. In response to this need, computing technologies are increasingly being applied to many aspects of manufacturing.
 Over the last three years, research activities in the Machine Intelligence Laboratory of HKUST have been centered on:
 design of openarchitecture machine tool controllers,
 development of algorithms for workpiece localization, and
 formulation in precise mathematical terms various geometric notions of tolerancing and study suitable approaches for tolerance verification.
 These technologies address important current manufacturing bottlenecks: the presently excessive costs of setup, refixturing and dimensional inspection. Our overall aim is to extend current onmachine probing inspection capabilities to reduce process time, improve machine tool flexibility, and increase part quality.
 In the present proposal, which supplements our on going laboratory work in an important way, we propose to study and analyze the geometric properties of two promising algorithms that have been developed by us and others for workpiece localization. These algorithms are the HT algorithm of Hong and Tan and the variational algorithm developed in one of our early works. We will use gradient flow techniques to study convergence properties and
 convergence rates of these two geometric algorithms. This study will lead to highly efficient localization algorithms which will possess the advantages of both the variational algorithm and the HT algorithm but none of their shortcomings. In parallel, we will develop envelopment algorithms for localizing partially finished workpieces as they proceed through successive stages of machining, and reliability analysis algorithms which can determine after an initial data acquisition step if tool paths can be positioned reliably enough relative to possible errors in the position of a part being cut, and if not can indicate where further probes should be made.
 We will also develop algorithms for automatic probing of threedimensional workpieces. By integrating these algorithms with an openarchitecture machine tool environment we propose to develop a ComputerAided Setup and Inspection (CASI) system. The use of a CASI system will eliminate the need of having an operator fixture a workpiece accurately, thus simplifying and accelerating the machining cycle. Furthermore, it will ensure product quality by performing onlinedimensional inspection.
Optimal Parallel Manipulator Design as Linear Matrix Inequality Problems
 Abstract
 This research deals with the problem of optimal design of parallel manipulators which have no singularity, have high stiffness and manipulability and are the most economic. By observing that those requirements can be cast into linear matrix inequalities (LMIs), we formulate the design problem as a convex optimization problem subject to LMIs with either a linear function or a maxdet function as its objective function. The variables x associated with LMIs are nonlinear functions of some key kinematic parameters /spl alpha/. If the dimension of x is equal to the number of independent kinematic parameters, a twolevel algorithm can be applied to solve for a set of optimal kinematic parameters:
 applying the interiorpoint algorithm for solving of x;
 applying the Newton method to a set of nonlinear algebraic equations for solving of /spl alpha/. If the dimension of x is greater than the number of independent kinematic parameters (i.e., x are not linearly independent), we consider the constrained semidefinite programming problems and the constrained maxdet problems by taking account of an additional set of nonlinear constraints. We propose a simplified constrained gradient algorithm for solving of x in such cases, /spl alpha/ derives from x using Newton method. Simulation results verify the effectiveness of the proposed algorithms.
A Geometric Theory for Analysis and Synthesis of sub6 DoF Parallel Manipulators
 Abstract
 This research developed a rigorous and precise geometric theory for the analysis and synthesis of sub6 DoF parallel manipulators. We give a rigorous definition for the parallel manipulator synthesis problem, and introduce a general method for specifying the corresponding subchains which will result in the desired parallel manipulator. Following this, a procedure for solving the parallel manipulator synthesis problem is proposed when the set of desired endeffector motions is in the form of Lie subgroup or a regular submanifold of SE(3). Numerous examples are used to illustrate the generality and effectiveness of the proposed synthesis method.
Parallel Manipulator Design for Optimal Workspace
 Abstract
 This work intends to deal with the optimal kinematic synthesis problem of parallel manipulators under a unified framework. Observing that regular (e.g., hyperrectangular) workspaces are desirable for most machines, we propose the concept of effective regular workspace, which reflects simultaneously requirements on the workspace shape and quality. The effectiveness of a workspace is characterized by the dexterity of the mechanism over every point in the workspace. Other performance indices, such as manipulability and stiffness, provide alternatives of dexterity characterization of workspace effectiveness. An optimal design problem, including constraints on actuated/passive joint limits and link interference, is then formulated to find the manipulator geometry that maximizes the effective regular workspace. This problem is a constrained nonlinear optimization problem without explicit analytical expression. Traditional gradient based approaches may have difficulty in searching the global optimum. The controlled random search technique, as reported robust and reliable, is used to obtain an numerical solution. The design procedure is demonstrated through examples of a Delta robot and a GoughStewart platform.
Students
H. Shen(Ph D) Winnie Leung(Ph D) G.F. Liu(Ph D) J.J. Xu(Ph D) Jie Li( Post doc) Y.S. Chan (RA) Tim Wu(Ph D)
K.T. Yuen(M Phil) D.J. Zhang(Ph D) Y.K. Yiu(Ph D) J. Meng(Ph D) Loginov A. (RA) Lin Shi (Junior RA) Toby (M Phil)
N. Chen(M Phil) Y. Lu(M Phil) Y.J. Lou(Ph D) A. Kojokarou(RA) KangKo A. (Post doc) H. Cheng(M Phil) Khrapko(RA)
Teaching
The list below is the courses that I have taught at HKUST. The course links will take you to the (current) course homepage.
 Elec 374: Introduction to Robotics
 Elec 564: Robot Manipulation
 Elec 560: Linear Systems Theory
 Elec 562: Nonlinear Systems
 Elec 112: Linear Circuit Theory
 Elec 692A: Special Topics in Robot Motion Planning
Contact
Mailing AddressZexiang Li 
Contact informationEmail: eezxli@ust.hk 
Other linksAutomation Technology Center 