There are a several kinds of force sensors normally referred to as torque cells that will measure torque and load cells that will measure force. Force transducers are devices that can measure torque directly within the mechanical system.
There are a number of reasons for direct measurement of force for the robot some of which include force quantitization, parameter optimization and weight measurement. Force transducers may be placed on the bipedal robot to determine the weight on each leg at any point in time.
Torque sensors may be fixed in robotic grippers in order to control gripper friction to ensure that nothing is crushed and anything that is picked is not dropped. The torque or force sensor can determine the moments and forces applied on a handle.
When compared to a layer of sensitive robotic skin covering the manipulator, the benefit of this kind of sensor is that it enables accurate measurement of shearing forces and applied moments instead of the normal forces.
Pires N et al submitted a paper on force/torque sensing applied to industrial robotic deburring with special emphasis on industrial robotic tasks. This study focused on a specific type of torque or force sensors, how these are used, and how they can be integrated into torque/force control applications using robots. The researchers also presented an industrial application wherein force control was mandatory for the task to succeed.
This research demonstrated an industrial force control application that was used for the deburring of metal parts coming from die cast in sand molds. The burr in the mold is very stiff making manual deburring tough and challenging for operators. Hence a robot was suggested. The robotic system included an industrial robot having a tool changer, a JR3 torque/force sensor, a grinding machine as well as a working platform for deburring.
This robot is designed to pick a piece for deburring, place it in the working platform or a rotating axis that has a tool changer for holding the piece that has to be deburred and a computer that makes the force control calculations and controls the robotic movements.
Lui T et al (2010) conducted research on reaction force/torque sensing in a master-slave robotic system without mechanical sensors. High precision and stability is needed in human–robot cooperative control systems. Traditional robot assistant systems normally implement force feedback with the help of torque or force sensors.
The researchers proposed an innovative force sensing mechanism that implemented force feedback in a master-slave robot system without any mechanical sensors. The system included two identical electro motors with the master motor overpowering the slave motor for interactions with the environment.
The novel force sensing mechanism was used to develop a bimanual coordinated training platform and verification of the system was done with experiments. According to results obtained, the proposed mechanism can achieve bilateral force sensing and mirror image movements of both terminals in opposite control directions.
Gu GM et al from the Department of mechanical engineering, KAIST proposed a single body torque sensor in 2011. Due to the simple single-body structure, the sensor has benefits of low cost and anti slip. Stress and strain analysis simulations were performed precisely. In order to demonstrate the performance of the proposed design, experiments were done that compared it with a commercial torque or force sensor.
Torque/force sensors are available commercially in a range of sizes with a number of measurement ranges. Input is provided by these sensors to the robotic controller, which makes use of this information to change its behavior conforming to a specific task.
A robotic arm can be either considered as a machine that positions and orients an end-effector or as a machine that enables the application of torque and force to the environment. There are a large number of robotic applications that strictly require torque/force feedback when compared to positioning measurements. Torque sensors in robots can be applied to the following:
- Force-Controlled Processes - There are a number of processes in manufacturing that need the application of an accurate force to attain quality results. For high-force operations like stamping, there will normally be dedicated machines. The use of a robot ensures high uniformity that will improve quality and bring down cost. In such applications, it is good to have a torque/force feedback to ensure high precision.
- Robotic Assembly - Insertion of a key into a keyhole may seem an easy task; however, it is actually coordination between our eyes and force as we wiggle the key in to make sure it enters right and does not damage the object into which it is being inserted. So in such cases the right fit for the insertion is a high-precision task for which visual feedback alone may not suffice and hence a better solution is a torque or force sensor.
- Telemanipulation - The controlling of robots from a distance by a human operator may need some force feedback. There may be delays in torque or force signals when tele-manipulation is done over a long distance.
- Human Augmentation - Human augmentation includes robotic prosthesis, enabling disabled people to perform daily activities. The intelligent assist device helps factory workers in the lifting and displacement of heavy objects. In this case, it is mostly necessary to determine the force applied by the operator for calculating an output velocity that will be sent as a command to the robot. The following video demonstrates application of torque sensors for force control in the iCub Humanoid robot:
Force control exploiting proximal force/torque sensors
SCHUNK has introduced the FTNet, which is an intelligent force and torque sensor that determines moments and forces in all six degrees of freedom. This sensor has an interface to the robot control unit that simplifies connection. The sensor is equipped with a high-speed data output of upto 7000 Hz for dynamic control. The sensor can be connected to a local area network
As electronics continues to become smaller and smaller and more electrically stable, it is possible to implement the sensors to higher spring rates that deliver improved dynamics of the measurement. This is achieved in such a way that at the same measuring accuracy, the measurement signals become small with the increased electric stability of the amplifiers.
It is, however, possible to implement the enhanced measurement signal conditioning for a high accuracy of the testing equipment layout. The future holds the advent of more intelligent sensors that will have stored measured technical data by which the measurements become highly secure and the data for quality assurance is called directly from the sensor.