Dec 18 2012
Image Credit: Willyam Bradberry/Shutterstock.com
Object recognition through touch enables the design and development of robots that can help humans in various fields. Robot technology finds applications in health and elder care, manufacturing, and high-risk environments. This involves handling objects made of different materials, with different shapes and in various positions relative to the robot. Advancements in sensing technologies allow robots to use touch and make sense of perceived information about objects before their interaction with it.
Tactile Sensing
Tactile sensing in humans is accomplished by mechanoreceptors. Mechanoreceptors are critical in stimulating the nerve impulses when the sensory neurons are exposed to the normal and shear forces, shape, softness, texture, temperature, pressure, of physical touch.
The key benefits of mechanoreceptors are as follows:
- They are capable of detecting body motion, sounds, and touch.
- They can monitor proprioception-sense of body position.
Mechanoreceptors come in two types- the Pacinian corpuscles and the muscle spindles. The Pacinian corpuscles present in the skin and other internal parts are linked to the sensory neurons for detecting mechanical pressure. Muscle spindles present within the muscle is activated with a response to a stretch reflex. Muscle spindles stimulate nerve impulses in the sensory neurons. The impulses are transmitted to the spinal cord to form synapses for coordinating various cellular activities within the human body human.
Tactile Transduction and Electronic Tactile Sensors – Research
The bio-inspired tactile sensor is provided with many microelectromechanical magnetic systems (MEMS) micro-force sensors. The sensors are covered by a soft layer containing a spherical elastomer cap that functions as derma-epiderma. A vertical silicon cylinder is in each micro-force sensor at its sensitive part and is fixed to a silicon membrane. The membrane consists of piezoresistive gauges through which the force applied on the cylinder can be obtained. A micro-positioning stage is provided for carrying the indentors and the substrates that are required to be scanned at the time of calibration.
The tangential and total normal loads can be measured using a double-cantilever system composed of capacitive position sensors. An application of a localized normal force on the surface of a soft layer at various positions accomplishes calibration of a tactile device.
Tactile Feedback and Prosthesis
Early studies conclude that object manipulation can be accomplished by feedback and feed-forward mechanisms. Meanwhile, individuals unable to use open-loop upper-limb prosthesis to receive tactile feedback may be due to insufficient grip force control and reduced dexterity. Saunders L et al (2011) experimented to identify if inadequate feed-forward or feedback control caused prosthesis control impairments.
As part of the experiment, a modified i-limb Pulse prosthetic hand, consisting of a differential force controller was fitted to healthy individuals. A vibrotactile feedback array is used to send grip-force feedback to their hands. Individuals were then directed to perform the tasks of grasping, lifting and replacing on a low-friction object. The following video demonstrates the functional principle of the i-Limb Pulse device by Touch Bionics.
The results showed that the grasp forces of the individuals were not within the linear range of the force sensor, which enabled them to find difficulties in grasping with the application of uncertainty on a hand controller to avoid feed-forward estimate. The grasp score was also compared for different feedback conditions.
A significant effect on tactile feedback was observed under conditions without visual feedback and no significant changes with tactile feedback. Therefore, it was concluded that the prostheses’s performance is lowered with the degradation of feed-forward predictability. The performance can be revived with the inclusion of tactile or visual feedback.
Improving Hand-Object Interaction
Fine controlled hand movements are executed through the combined work of multiple elements. This process is defined as synergy. Robotics have successfully utilized the synergy framework. The Hand Embodied (THE) is a project that integrates the fields of robotics and neuroscience to replicate the teamwork model of movement execution.
The control of the hands is accomplished through spatial-temporal coordination. Researchers from Yale University’s GRAB Lab have utilized a variable-friction system to develop a two-finger design for hand manipulation. This is inspired by the biomechanical properties of human fingers. As the force needed act is increased, the friction of the fingers changes to accommodate the workload when choosing whether to grip or slide over objects.
The structure of the hands allows the performance of such controlled hand movements. The bone structure of the hands allows the application of gripping force when desired. The skin helps us maintain softer contact with objects. By using friction contact alongside softer contact, we can perform all kinds of movements. The GRAB Lab’s variable friction fingers replicate human finger functionality. This design accomplishes that by turning friction on and off to accommodate the desired action: gripping or sliding. This simple, yet clever design can be a foundation for creating even more complex robotics aiming to improve hand-object interaction. These robotics can aid the rehabilitation of patients with motor disabilities by providing an alternative, yet seemingly more natural interaction.
Future Research and Challenges
The following are some of the challenges involved in tactile sensing research:
- Frequency response - It is evident from research that sensors having a minimum frequency response of 32Hz are suitable for identifying incipient slips in prosthetic applications. Likewise, sensors more appropriate for the detection of vibrations at low spatial resolution normally provide a frequency response of 250 Hz. Though several PVDF-based sensors are available, there still exists a need for a sensor that detects static forces.
- Task-centered design criteria - Specifications of task-based design criteria can offer more advanced solutions in the field of robotics and biomedical technology which reduces the cost of the sensor.
- Spatial resolution – Studies suggested that around 1.25mm spatial resolution is required for measuring an innervation density of mechanoreceptors. Hence, sensors having greater spatial resolution and greater individual points scanning frequency are required.
- Assembly and maintenance - The assembly and disassembly of sensors have to be convenient, especially for those that find application in medical surgery where disposable equipment is used.
- Cost – Inexpensive tactile sensors with low hysteresis and high wearability and repeatability are in great demand as surgeries like MIS mostly make use of disposable instruments.
- Arrayed sensor design and algorithms – Though well-known techniques like optical, piezoelectric, resistive and capacitive techniques exist, algorithms for characterization/discrimination and customizable interfaces are still needed.
Sources and Further Reading
- Candelier R, Prevost A, Debr´egeas G. The Role of Exploratory Conditions in Bio-Inspired Tactile Sensing of Single Topogical Features. Sensors. 2011:11;7934–7953.
- Saunders L, Vijayakumar S. The role of feed-forward and feedback processes for closed-loop prosthesis control. Journal of NeuroEngineering and Rehabilitation. 2011;8(60).
- Tiwana M.I, Redmond S.J, Lovell N.H. A review of tactile sensing technologies with applications in biomedical engineering. Elsevier. 2012:17–31.
This article was updated on 12th February, 2020.