RRG Research Force/Motion Control

Research

Force/Motion Control of Intelligent Mechanical Systems

Motivation

In most cases, when a manipulation device performs a contact task, the force experienced by the effector has to be managed (or controlled) in conjunction with controlling its motion. This is due to the fact that the effector is dynamically interacting with the environment. For this reason, such a control or decision making scheme can be called Interaction Control or Force/Motion Control (FMC)

Objective

In the Force Control research thread, UTRRG's objectives are

To conduct fundamental research and develop key technologies in three principal areas: Model-Based Process Control, Component Level Technology and Operational Software
To apply the results from this research to serial robotic systems and demonstrate their extensibility to other similar nonlinear coupled mechanical systems (Parallel Robots, Walking Machines, Multi-Fingered Hands etc)

Approach

Model-Based Process Control: The concept of force/motion managament based on a performance map/model of the system based on as-built parameters is termed as Process Control. Any form of force/motion controller may be implemented as part of this paradigm. As regards force/motion control algorithms, the two main research threads are impedance control and hybrid force/position control. These architectures will be empirically compared using performance metrics from representative task scenarios. Cases where force and motion are managed in mutually exclusive directions will be considered in addition to ones where these are managed in the same sub-space.

Component Level Technology: To facilitate fault-tolerance, UTRRG has developed an array of multi-input actuators, both velocity summing and force summing. On the component level we are now investigating the possibility of embedding multi-domain input in the same actuator (Force/Motion Actuator). In such an actuator, the force and motion subsystems have a scale change of 10 to 15-to-1. The influence of such a actuator on the choices in output/task space and also the mixing of force/motion criteria are issues under consideration.

Operational Software: An objective of this technical research thread at the system level is to develop a domain/sub-domain within OSCAR which is modular, extensible, reusable, robot-independent and simplifies application development for modeling, sensing and control of real robotic systems and simulation models. To this end a control systems domain is being developed within which the force control components will be implemented.

Back to Top

Force Control Testbed

System: A modular manipulator testbed is used for all force control experiments. It consists of the Powercube^TM actuator modules manufactured by Amtec GmbH, Germany. These modules come with standardized interfaces to facilitate reconfiguration. The actuator modules can be controlled in position, velocity and current modes and use the CAN bus for communication.

Figure 1. Testbed Components (Left: PowerCube Manipulator, Right: ATI F/T 30/100 Sensor)

Sensing: To sense contact forces, a wrist-mounted multi-axis ATI FT 30/100 Force/Torque (F/T) sensor is used. This measures all six components of the end-effector forces and torques. The axial, transverse and torsional load capacities are 300 lbs, 150 lbs and 600 lb-in respectively. The maximum attainable data transfer rate is 600 Hz with the controlled system. Accuracy (including manipulator vibration) during operation is 0.25 lbs.

Software: The operational software used for developing force control applications is OSCAR

Research Results/Demonstrations

Human Augmentation Technology (HAT)

By means of force/motion scaling using mechanical systems, it is possible to augment (or at least supplement) the capabilities of an operator. This technology is called Human Augmentation and may be used in a variety of application domains ranging from battlefield operations (similar to exoskeletons) to stroke patient therapy/rehabiliation.

HAT Demonstrations

The following demonstrations are based on HAT and have been implemented using our in-house operational software (OSCAR). The same applications have been tested on two different types of manipulators (modular and monolithic systems) thus demonstrating the generality and extensibility of our software. The modular system used is the powercubes from Amtec and the monolithic system is the FANUC M6iB-6S.

Payload Assist Demo (PowerCube Robot) : In this demo the modular manipulator's (Powercube) end-effector behaves like an inertia. This could be used to do force-scaling for human augmentation. This same application can be used to create a master manipulator for a teleoperation system. End-effector forces are mapped to differential motions in cartesian directions. This application also aids in ensuring the correctness of the frame transformations from sensor frame to the task/world frame and can thus be used to verify frame transformations for a force-controlled application.

Payload Assist Demo (FANUC M6iB-6S) [~4MB] : In this demo the above application was implemented on a standard commercially available industrial robot (FANUC M6iB-6S).

Virtual Spring Demo : In this demo the modular manipulator's (Powercube) end-effector emulates a spring. A delta controller was used to develop this application. The advantage of such programmable compliance is that we could change the behaviour of the end-effector on the fly depending on the task. The compliances can be different in different directions too. Also, the dynamic behavior of the end-effector can now be programmed as a combination of the above two modes, inertia and spring.

Rehabilitation Demo (~21 MB, 1:07 min) : In this demo the above HAT application was reconfigured so that a modular manipulator could assist a therapist in teaching a pre-planned path. Robots are increasingly being used for therapeutic applications. Once the path has been taught, the manipulator can be run at various speeds depending on the therapy needs. Such an application helps stroke patients regain their motor functionality.

Contact Control

Transition from free space motion to constrained space motion can excite undesirable bouncing actions depending on the approach velocities and surface characteristics. The strategy followed for contact control was to execute guarded motion immediately after contact is established. Contact is sensed by monitoring the force signal derivative, after ensuring that this signal is 'clean'. A velocity based control was implemented using the following law.

Contact Control Demonstrations

Force Regulation Demo : A force regulation task (target force = 3 lbs) with a compliant environment is shown in this video. The approach velocity used was 5 mm/s.

Active Force Control

Most industrial robots are used in the position control mode with little knowledge about the forces of interaction between the manipulator and its environment. However, the need to incorporate force sensing and active force control is being felt in the robot industry and force control has already been implemented for some applications such as assembly and deburring. The following videos show some demonstrations we developed using OSCAR to implement force-compliant mechanical finishing. The integrated system consists of a ATI F/T sensor, FANUC M6iB-6S industrial robot, and a stepper-motor controlled wire brush. The software libraries used to develop this application were Open Motion API from FANUC and OSCAR by UTRRG.

Active Force Control Demonstrations

Force Controlled Finishing of Planar Part [~23MB] : This movie demonstrates a force controlled finishing operation using the FANUC M6iB-6S industrial manipulator operating on a steel planar surface using a steel wirebrush. The controller used in this demo was a PD force controller.

Force Controlled Finishing of Spherical Part [~13 MB] : This movie demonstrates a force controlled finishing operation using the FANUC M6iB industrial manipulator operating on a sphericalsteel surface using a steel wirebrush. The controller used in this demo was a PD force controller. Following are the links to movies that show the GUI an operator would use to teach trajectory points, input force controller parameters, launch the force controller application, and display the process statistics.

Force Controller GUI: Process Setup [~8 MB]

Force Controller GUI: Process Statistics [~2 MB]

Generalized Impedance Control of Redundant Manipulators

Impedance Control achieves force control and motion tracking in the same direction by specifying a programmable dynamic response between the manipulator end-effector and the environment. For details, refer [Pholsiri et al, 2003] in the publications section.

Impedance Control Simulations

Compliant Control Saw-Cutting : In this demo a 10DOF redundant manipulator is used to execute a saw cutting task while optimizing one or more performance criteria.

Saw-Cutting while Minimizing Force Transmission Ratio (FTR)

Saw-Cutting while Maximizing Measure of Transmissibility (MOT)

Publications

Pholsiri. C, Rabindran. D, Pryor. M, and Kapoor. C, December 2003, "Extended Generalized Impedance Control for Redundant Manipulators", Proceedings of the IEEE Conference on Decision and Control , Hawaii [Abstract] [Full-text PDF]

Cox. D. J, Pryor. M, Cetin. M, and Tesar. D, June 1999, "Experiments in Cooperative Manipulation for Dual Arm Robotic Operations", Proceedings of the 10th World Congress on Theory of Machines and Mechanisms [Abstract] [Full-text PDF]

Hester. R. D, Cetin. M, Kapoor. C, and Tesar. D, May 1999, "A Criteria-Based Approach to Grasp Synthesis", Proceedings of the IEEE Conference on Robotics and Automation , Vol. 2, pp. 1255-1260 [Abstract] [Full-text PDF]