A Neural Network Approach to the Frictionless Grasping Problem
01 Sep 2000-Journal of Intelligent and Robotic Systems (Kluwer Academic Publishers)-Vol. 29, Iss: 1, pp 27-45
TL;DR: This article presents a heuristic technique used for solving linear complementarity problems (LCP) that finds almost exact solutions in solvable positions, and very good solutions for positions that Lemke fails to find solutions.
Abstract: This article presents a heuristic technique used for solving linear complementarity problems(LCP). Determination of minimum forces needed to firmly grasp an object by a multifingered robot gripper for different external force and finger positions is our proposed application. The contact type is assumed to be frictionless. The interaction in the gripper–object system is formulated as an LCP. A numerical algorithm (Lemke) is used to solve the problem [3]. Lemke is a direct deterministic method that finds exact solutions under some constraints. Our proposed neural network technique finds almost exact solutions in solvable positions, and very good solutions for positions that Lemke fails to find solutions. A new adaptive technique is used for training the neural network and it is compared with the standard technique. Mathematical analysis for the convergence of the proposed technique is presented.
Citations
More filters
TL;DR: A two stage method for constructing a firm grip that can tolerate small slips of the fingertips and is a robust, reliable, and stable controller for rigid bodies that can be handled by a robot gripper.
Abstract: This paper presents a two stage method for constructing a firm grip that can tolerate small slips of the fingertips. The fingers are assumed to be of frictionless contact type. The first stage was to formulate the interaction in the gripper–object system as a linear complementarity problem (LCP). Then it was solved using a special neural network to find minimal fingers forces. The second stage was to use the obtained results in the first stage as a static mapping in training another neural network. The second neural network training included emulating the slips by random noise in the form of changes in the positions of the contact points relative to the reference coordinate system. This noisy training increased robustness against unexpected changes in fingers positions. Genetic algorithms were used in training the second neural network as global optimization techniques. The resulting neural network is a robust, reliable, and stable controller for rigid bodies that can be handled by a robot gripper. © 2001 John Wiley & Sons, Inc.
13 citations
TL;DR: The importance of providing lean‐flexibility by means of reconfigurable, automated robot hand changers (ARHC), particularly in small‐batch robotic welding, assembly, machine loading and in other flexible robot cells, is discussed with examples.
Abstract: Robot tools, or in more general terms, end‐of‐arm tools, or robot end‐effectors are general purpose, programmable or task‐oriented devices connected between the robot wrist and the object or load to be manipulated and/or processed by the robot. They can offer and/or limit the versatility of grasping and/or processing of different components, sensing their characteristics and working together with the robot control system to provide a reliable “service” throughout the component manipulation cycle. Reconfigurable robot tooling enables the robot to rapidly change its end‐effectors or fingers of its end‐effectors, typically under programmable software control. The importance of providing lean‐flexibility by means of reconfigurable, automated robot hand changers (ARHC), particularly in small‐batch robotic welding, assembly, machine loading and in other flexible robot cells, is discussed with examples. Some known systems are demonstrated and the “Ranky‐type” ARHC design is illustrated in more detail.
12 citations
Cites background from "A Neural Network Approach to the Fr..."
...(Please note, that any of the listed aspects could identify different groups of robot tools, or single, or multi-task grippers for classification and system integration purposes) ( Abu-Zitar and Al-Fahed Nuseirat, 2000...
[...]
TL;DR: The results show that ACO can find optimum fingers forces for grasping rigid objects and the optimal set of parameters used to tune ACO is independent of the initial number of ants on each location.
Abstract: – The problem of estimating the minimum forces extracted by robot fingers on the surface of a grasped rigid object is very crucial to guarantee the stability of the grip without causing defect or damage to the grasped object. Solving this problem is investigated in this paper. Moreover, the optimum sets of parameters used to tune the algorithm are also studied here., – Ant Colony Optimization (ACO), which is a swarm intelligence‐based method, is used in this work to solve this problem. The problem under scope is a complex, constraint optimization problem. We develop our own approach to calculate those minimum forces. Ants ability to reorganize and behave collectively is modelled here. The required forces are a result of the final ants distribution around the fingers contact points. Ants move from contact point to another following the maximum pheromone level direction until they settle on a solution that accomplishes the given criteria. Ants number on a contact point constitutes the total force exerted by a finger on that contact point. The process is repeated until optimum solution is found. Simulations are repeated to track down most suitable ACO parameters for this type of problems and with different fingers configurations., – The results show that ACO can find optimum fingers forces for grasping rigid objects. These objects could be any polygon with or without friction between the fingers tips and the object surface. The method is computationally acceptable and can be applied with different fingers configurations and with different friction coefficients. We found that the optimal set of parameters used to tune ACO is independent of the initial number of ants on each location., – In this paper we present a very original, new, and interesting technique used to solve the optimum grasping forces of rigid objects. It is a well‐known fact that standard optimization techniques have their own requirements and limitations. This technique is based on swarm intelligence. This work opens the door for further investigations on how nature based methods can be used to solve complex problems. ACO offers a simple, yet structure approach to solve nonlinear constraint optimization problems.
12 citations
Cites methods from "A Neural Network Approach to the Fr..."
...) or it could be solved using Neural Network based methods ( Abu-Zitar and Al-Fahed Nuseirat, 2000...
[...]
TL;DR: An intelligent rule-based method that figures out the minimal number of fingers and minimal values of contact forces required to securely grasp a rigid body in the presence of friction and under the action of some external force is proposed.
Abstract: This paper presents an approach for grasp planning and grasp forces optimization of polygon shaped objects. The proposed approach is an intelligent rule-based method that figures out the minimal number of fingers and minimal values of contact forces. These fingers are required to securely grasp a rigid body in the presence of friction and under the action of some external force. This is accomplished by finding optimal contact points on the object boundary along with minimal number of fingers required for achieving the aforementioned goal. Our system handles every object case independently. It generates a rule base for each object based on adequate values of external forces. The system uses the genetic algorithm as its search mechanism, and a rule evaluation mechanism called bucket brigade for the reinforcement learning of the rules. The process mainly consists of two stages; learning then retrieval. Retrievals act on line utilizing previous knowledge and experience embedded in a rule base. If retrievals fail in some cases, learning is presumed until that case is resolved. The algorithm is very general and can be adapted for interface with any object shape. The resulting rule base varies in size according to the degree of difficulty and dimensionality of the grasping problem.
11 citations
Cites methods from "A Neural Network Approach to the Fr..."
...Linear programming techniques as well as heuristic search methods were used in finding the best solutions [10, 11]....
[...]
...gorithms (GA’s), or even Adaptive Linear Mean Square Methods (ALMS) have been successfully used in solving this problem [11]....
[...]
Dissertation•
01 Jan 2013
TL;DR: The research work aims at developing an improved anthropomorphic robot hand model in which apart from the four fingers and a thumb, the palm arch effect of human hand is also considered to increase its dexterity.
Abstract: The ability of stable grasping and fine manipulation with the multi-fingered robot hand with required precision and dexterity is playing an increasingly important role in the applications like service robots, rehabilitation, humanoid robots, entertainment robots, industries etc A number of multi-fingered robotic hands have been developed by various researchers in the past The distinct advantages of a multi-fingered robot hand having structural similarity with human hand motivate the need for an anthropomorphic robot hand Such a hand provides a promising base for supplanting human hand in execution of tedious, complicated and dangerous tasks, especially in situations such as manufacturing, space, undersea etc These can also be used in orthopaedic rehabilitation of humans for improving the quality of the life of people having orthopedically and neurological disabilities The developments so far are mostly driven by the application requirements There are a number of bottlenecks with industrial grippers as regards to the stability of grasping objects of irregular geometries or complex manipulation operations A multi-fingered robot hand can be made to mimic the movements of a human hand The present piece of research work attempts to conceptualize and design a multi-fingered, anthropomorphic robot hand by structurally imitating the human hand
In the beginning, a brief idea about the history, types of robotic hands and application of multi-fingered hands in various fields are presented A review of literature based on different aspects of the multi-fingered hand like structure, control, optimization, gasping etc is made Some of the important and more relevant literatures are elaborately discussed and a brief analysis is made on the outcomes and shortfalls with respect to multi-fingered hands Based on the analysis of the review of literature, the research work aims at developing an improved anthropomorphic robot hand model in which apart from the four fingers and a thumb, the palm arch effect of human hand is also considered to increase its dexterity
A robotic hand with five anthropomorphic fingers including the thumb and palm arch effect having 25 degrees-of-freedom in all is investigated in the present work Each individual finger is considered as an open loop kinematic chain and each finger segment is considered as a link of the manipulator The wrist of the hand is considered as a fixed point
The kinematic analyses of the model for both forward kinematics and inverse kinematic are carried out The trajectories of the tip positions of the thumb and the fingers with respect to local coordinate system are determined and plotted This gives the extreme position of the fingertips which is obtained from the forward kinematic solution with the help of MATLAB Similarly, varying all the joint
iv
angles of the thumb and fingers in their respective ranges, the reachable workspace of the hand model is obtained Adaptive Neuro-Fuzzy Inference System (ANFIS) is used for solving the inverse kinematic problem of the fingers
Since the multi-fingered hand grasps the object mainly through its fingertips and the manipulation of the object is facilitated by the fingers due to their dexterity, the grasp is considered to be force-closure grasp The grasping theory and different types of contacts between the fingertip and object are presented and the conditions for stable and equilibrium grasp are elaborately discussed The proposed hand model is simulated to grasp five different shaped objects with equal base dimension and height The forces applied on the fingertip during grasping are calculated The hand model is also analysed using ANSYS to evaluate the stresses being developed at various points in the thumb and fingers This analysis was made for the hand considering two different hand materials ie aluminium alloy and structural steel
The solution obtained from the forward kinematic analysis of the hand determines the maximum size for differently shaped objects while the solution to the inverse kinematic problem indicates the configurations of the thumb and the fingers inside the workspace of the hand The solutions are predicted in which all joint angles are within their respective ranges
The results of the stress analysis of the hand model show that the structure of the fingers and the hand as a whole is capable of handling the selected objects
The robot hand under investigation can be realized and can be a very useful tool for many critical areas such as fine manipulation of objects, combating orthopaedic or neurological impediments, service robotics, entertainment robotics etc
The dissertation concludes with a summary of the contribution and the scope of further work
6 citations
Cites methods from "A Neural Network Approach to the Fr..."
...Zitar and Nuseirat [155] using inequality theory the new neural network architecture to solve the arisen linear complementarity problems was developed....
[...]
References
More filters
17 Jul 1986
15,313 citations
9,153 citations
Book•
01 Jan 2009
TL;DR: The aim of this book is to provide a Discussion of Constrained Optimization and its Applications to Linear Programming and Other Optimization Problems.
Abstract: Preface Table of Notation Part 1: Unconstrained Optimization Introduction Structure of Methods Newton-like Methods Conjugate Direction Methods Restricted Step Methods Sums of Squares and Nonlinear Equations Part 2: Constrained Optimization Introduction Linear Programming The Theory of Constrained Optimization Quadratic Programming General Linearly Constrained Optimization Nonlinear Programming Other Optimization Problems Non-Smooth Optimization References Subject Index.
7,278 citations
"A Neural Network Approach to the Fr..." refers background in this paper
...In our case,F is positive-definite (PD), so the solution r is also unique [8]....
[...]
TL;DR: Results of computer simulations of a network designed to solve a difficult but well-defined optimization problem-the Traveling-Salesman Problem-are presented and used to illustrate the computational power of the networks.
Abstract: Highly-interconnected networks of nonlinear analog neurons are shown to be extremely effective in computing. The networks can rapidly provide a collectively-computed solution (a digital output) to a problem on the basis of analog input information. The problems to be solved must be formulated in terms of desired optima, often subject to constraints. The general principles involved in constructing networks to solve specific problems are discussed. Results of computer simulations of a network designed to solve a difficult but well-defined optimization problem-the Traveling-Salesman Problem-are presented and used to illustrate the computational power of the networks. Good solutions to this problem are collectively computed within an elapsed time of only a few neural time constants. The effectiveness of the computation involves both the nonlinear analog response of the neurons and the large connectivity among them. Dedicated networks of biological or microelectronic neurons could provide the computational capabilities described for a wide class of problems having combinatorial complexity. The power and speed naturally displayed by such collective networks may contribute to the effectiveness of biological information processing.
5,328 citations
01 Jan 1988
3,613 citations
"A Neural Network Approach to the Fr..." refers methods in this paper
...(31) In the ALMS,ρk is a random variable taken from the normal distribution with σk = g(Ek), (32) whereσ is the standard deviation,E is the value of MSE, andg(Ek) is a linear function....
[...]
...It is clear from Figure 9 that our ALMS algorithm has outperformed the standard LMS algorithm especially at the last stages of learning....
[...]
...The ALMS was faster in approaching the low error zone than did the standard LMS as shown in Figure 9....
[...]
...Our ALMS algorithm, however, adaptively modifies the learning rate, in a manner, maintains larger search steps in the regions with a high error energy, and zooms down the learning rate as the energy error goes down....
[...]
...This behavior is due to the random tools that we used in the ALMS algorithm to prevent getting stuck in flat regions or in oscillations [5]....
[...]