biometric robot for seminar

 Biomimetic Robots

 ABSTRACT

 

A lot of conventional thinking is prevalent in the field of robotics as a whole and this is an attempt to present a brief idea of things being done by researchers throughout the world in the field of biomimetic robots. Biomimetic or bioinspired robots is a broad term that applies to those robots whose design has been created or copied from the biological data of various organisms. It is an attempt by the researchers world-wide to understand how birds fly, how dolphins swim in water, how humans walk and apply those activities in a device which can lead to better technologies for the society in general.  A brief introduction into the various research projects around the world has been given. Research groups have hopes of developing robots that might possibly detect mines, explore Mars through BEES (Bioinspired engineering of exploration systems), or even search for trapped people in disaster situation.  

 1.INTRODUCTION

What is biomimetic? Biomimetic is a general description for engineering a process or system that mimics biology. The term has emerged from the field of biochemistry and applies to an infinite range of chemical and mechanical phenomenon, from cellular processes to whole-organism functions. The field itself is relatively new and collaborations required between biologists, robotics engineers and computer science engineers have barely begun.

Mark Cutkosky, a professor in Stanford University’s Department of Mechanical Engineering defines it as “ extracting principles from the field of biology and then applying it to man-made devices – particularly robots.” Talking of the programming being done by the various research groups, Butler Hine, manager of the Computing information and communications technology program based at NASA- Ames, believes that “biomimetics uses nature’s evolved way of doing things rather than the computationally intensive way”.

Biomimetics in a broad sense is a convergence of various fields of science as a whole and provides a chance for looking into the living systems for inspiration for creation of new robots. It requires the expertise of a team of investigators from diverse backgrounds to understand the various complex processes involved. For example, the movements made by animals or legged robots are generated by actions within a control system ultimately causing a mechanical system to act upon objects. In animals, control is found in the nervous system and the muscles and skeleton make up the mechanical part.

Four view points are to be considered over here to get the exact output.

-          First is a biological viewpoint wherein biologists understand at a neurobiological and behavioral level how animals move through their environment.

-          Second, are those engineers who develop the control architectures through computer modeling.

-          Third, biomedical engineers who develop prospective neural network devices that would interface with the mechanical systems to regulate movements.

-          Fourth is a mechanical engineering perspective that seeks to construct artificial systems that move through complex terrain and manipulate objects in assembly.

The diversity found in the goals and backgrounds of each of the four individual groups combines to permit an investigation of biomechanical system. A common goal here would be to understand how movement results from an intimate interaction between control systems that direct movement and mechanical systems that execute it.

Various research groups around the world are working on various different projects out of which a brief overview of some I found really interesting is presented here. Projects varying from land, water and air have been discussed.

2.UNDERWATER ROBOTICS

Fig. 1 Joseph Ayers, along with his team at NorthEastern University are carrying out research on the field on robotic lobsters. Particularly, they are testing the locomotion characteristics

Two research groups, one based in Brooklyn College and headed by Frank Grasso and another based in Northeastern University and headed by Joseph Ayers, have been working on different aspects of the lobsters. The first is working on sensory-chemical tracking and lobster’s acute sense of smell to sense prey in all circumstances while the latter team is working on locomotion systems that would probably be the first robot with an elementary nervous system. The robots would be using neuronal circuit based controllers, artificial muscle based on smart materials and microelectronic sensors that code in the same fashion as animal sensors.

Frank Grasso, associate professor of psychology at Brroklyn College[9][8], is working on lobster’s ability to sense its prey in turbulent ocean conditions. Work on Robolobsters I and II has been completed and they match the size as well as the sensing and locomotive abilities of their biological counterparts. These robots let the researchers test the hypotheses based on the lobster’s ability to detect olfactory information and make decisions based on it. As working of these robots is primarily under water, it is not possible to have it attached to power source.

Joseph Ayers, a biology professor at Northeastern University[8][6], is working on the locomotion system of the lobsters. Typically in the field of robotics, motors are used to achieve locomotion but he is trying to use neurons to analog VLSI that would be subsequently used to generate “motor program-like central pattern generators based on neurons”. The whole working has been divided into two components: the central system and the peripheral system. Central system consists of the following major components:

- Segmental central pattern generators (CPGs) that control the motor neurons and ultimately the muscles of each limb.

- Coordinating systems that determine the phase relations or gaits between the CPGs of different limbs.

- Command Systems that specify and modulate the behavior generated by the CPGs. The command systems represent the control locus at which the decision to generate a particular behavior is achieved.

In addition to this, peripheral systems are designed consisting of sensors that provide feedback to the central pattern generators.

- Exteroceptive or Orientational Reflexes that operate at the level of the command systems to generate whole-body compensatory responses.

- Phase Modulating Reflexes that operate at the CPG level to reset the timing of oscillations during stumbles, etc.

- Amplitude Modulating Reflexes that operate at the motor neuron level to control the amplitude of the motor output.

One of the major problems faced by the research group is that of an actuator that approximates muscle. Crustacean muscle joints has mechanical properties that are quite different from typical robotic actuators and these properties need to modeled to achieve interface between controlling neuronal signals and realization of joint angular movements. This is a very difficult problem considering the fact that crustacean muscle is typically modeled as a tension generating component (the contractile fibers) in series with an elastic component (tendons), both of which are in parallel with an elastic component (muscle cell membranes). Neuromuscular control, on the other hand, is mediated by motor neurons that differ in the number of muscle fibers that they recruit.

DARPA, ONR and NSF are primarily funding both the researches. However, the lack of a single goal to find an appropriate use of the technology is the only hindrance in achieving the required results. Joel L. Davis, an ONR professor working on adaptive neural systems expects that one robotic lobster should be ready for use by summer 2006.

These robots can fulfill a variety of missions including remote sensing, ship tagging, and mine countermeasures. Deep underwater surveillance can also be carried out using these devices. A military application as discussed would be to search for mines on beachheads and they can be further designed to work underwater or on the surface of land in shallow beach areas. If one lobster detects a mine, for example, it should signal others in the area to go away before it detonates it. These applications are at a very preliminary stage and require lot of planning before reaching the implementation level.

3.BIOMIMETICS IN THE AIR

The Biorobotics Research Group[1]at Movement and Perception Lab, France have been working on biologically inspired aerial vehicles for quite some time and have been successful enough to develop systems that can be mounted on-board Micro-Air-Vehicles. They have developed two Automatic Flight Control Systems (AFCS): OCTAVE (Optical altitude Control system for Autonomous Vehicles) and OSCAR (Optical Scanning sensor for Autonomous Robots).

The whole system is primarily based on EMD (Elementary Motion Detectors), which forms the visual processing system of the OCTAVE and OSCAR. The system consists of two photoreceptors Ph1 and Ph2, the visual axes of which are separated by an interreceptor angle. The scheme of EMD looks as follows:

1.      Band-pass filtering on each channel.

2.      Thresholding and pulse generation on each channel.

3.      Generating a long-lived decaying signal on one channel.

4.      Generating a very short, unitary sampling pulse on the other channel.

5.      Minimum-detection based sampling on the other channel delievers a signal that decreases monotonically with time.

The OCTAVE system works mainly on concept of relative motion of an aerial vehicle with respect to the contrasting terrain underneath. The OCTAVE robot can lift itself automatically so as to follow a terrain over a wide range of ground speeds, by essentially relying on the data provided by EMD sensor that is pointed downwards.

The OSCAR sensor, on the other hand, responds to the micro-scanning movements of the visual sensor itself. The system based on the micro-scanning system in the flying fly can be used for detecting and locating targets to an accuracy of fraction of pixel period.

At Georgia Tech Research Institute[5], research work is being done on a device called the

‘Entomopter’. The generation of the Entomopter is designed for operation in two different atmospheres: a 50-gram terrestrial version and an aerospace version designed for use in different gravitational environments. Both of them are made from carbon composite materials. The design process is such that it is a circulation control process that turns high speed, hot-gas flow into lower speed, cooler gas that can, when vented out through the wing, cause flow that gives the vehicle seven times more lift and lets it fly at lower speeds. However, flight control processes will start only after flight mechanics is solved. This project is being funded by DARPA and NASA too is interested on deploying such a vehicle for Mars mission.

Though initially it was planned that this micro-sized air vehicle would go over a hill or find an obstacle and find enemy location, it soon proved to be impractical for problems related to line-of-sight communication. So, the group has since shifted focus to indoor operations wherein the devices are nimble enough to enter a building through a chimney or open window, fly fast to evade camera detection and carry out surveillance mission[8].

4. BIOMIMETICS ON LAND

Stanford’s Centre for Design Research[8,3,4], has a team that is designing and fabricating six-legged robots that try to resemble the physical construction and mechanical design that are responsible for the robustness of the cockroach. Points that have been difficult to understand the mechanism of the cockroach are several.

Cutkosky notes that “ They run over obstacles without slowing down or getting knocked off course, and they do this mainly by virtue of having a wonderful tuned mechanical system-sort of suspension of car- that keeps then stable and on course. It’s hard to damage a cockroach.”

The Sprawl robots, which the team is presently designing, are unique in the sense that they rely on good mechanical design rather than faster computational properties. The various aspects considered in the design of the robot are:

1.Self-stabilizing posture.

2. Thrusting and stabilizing leg function.

3. Passive visco-elastic structure.

4. Timed, open-loop/feed forward control.

5. Integrated construction.

The robot is able to move autonomously up to 15 body lengths per second. The system has six legs that move in alternating –tripod gait in a sprawled design that mimics the cockroach’s biological structure and allows the robot to move faster. Shape deposition manufacturing process has been used for embedding the various sensors, actuators and microprocessors. It is a complex process that has given rise to a durable polymer shell. 

Another robot that is inspired by the cockroach is the Robot Hexapod Rhex (pronounced as “rex” )[2] is considered to be the most maneuverable robot built. It’s about the size of a shoebox. It can bump along at up to 3 meters per second and last close to two hours on one battery charge. It has

Fig.6 The Whegs, are being developed at the Case Western Reserve University and are inspired by a hybrid mechanism of wheels and legs.

six legs, which are capable of clockwise rotation. It’s a possible candidate for the 2007 mission to Mars, which needs surface robots to navigate rough terrain. It uses a total of 6 motors and spoke-like legs.

The Wheg series of robots beings developed at Case Western Reserve University[7] use insect behavior study and other biological data. The term ‘whegs’ is derived from the words ‘WHEels’ and ‘leGS’. It utilizes a method of locomotion that combines the advantages of wheels and legs. Wheels are easier to use and allow vehicles to move faster over terrain quickly. Legs allow robots to climb obstacles that are higher than what a wheeled vehicle would be able to climb over. Whegs use just one motor for propulsion, which gives it the advantage of being able to supply to any leg with the maximum power available on board the vehicle.  If individual motors are used to drive each leg, then each must be powerful enough for bad scenario. Here, same amount of torque can be applied by much less weight of motor.

IRobot has built an autonomous robot, Ariel[2], on the concept of crab, which is capable of moving on its own. It is probably the first legged robot capable of walking either on land or underwater in the turbulent surf zone. It can climb over obstacles and crevices that would block traditional wheeled vehicles and can also resist the impact of waves. Ariel is also completely invertible — if flipped over by the waves, its legs simply reorient so that the "top" of the body becomes the bottom. It is undergoing tests in which it locates mines in the surf zone.

Shape Deposition Manufacturing process[3] is a layered prototyping method where parts or assemblies are built up through a cycle of alternating layers of structural and support materials. After a layer of material is added, it is then shaped to a precise contour before the next layer is added. The intermittent addition of support material allows for the construction of arbitrary geometries. Unlike many other layered processes, the material is shaped after it is added. This allows for high precision.

Figure 8 describes the construction of multi-material compliant leg that was used in the robots. It takes the advantage of SDM’s capability to vary the material properties during construction of the part. Each layer was built up on a different material, each with its own characteristics. The deposition of a layer of viscoelastic polyurethane creates the compliant, damped hip flexure joint. A stiffer grade of polyurethane had been used for structural members, which encase the piston and servo mounting.

inside the part in a precise and repeatable fashion was used by them to create the robot’s body. As shown, the robot’s servos, wiring and connectors were embedded within the body. They did this by first shaping the support substrate (high melting-temperature wax) as a mold for the bottom of the body. The embedded components were then placed and protected by sacrificial material (low melting-temperature wax). A layer of structural material (pourable polyurethane) was then deposited and shaped, thereby encasing the embedded components. Finally, the sacrificial material was removed to access the finished part.

All in all, this method of manufacturing process in the field of robotics can definitely help in improving the precision with which the work is done. Especially in the case of small robots, it is generally quite difficult to manage the joints and fixtures and so fasteners are used. However, in such situations, use of this kind of manufacturing practices can add great value to the overall time saved.

Fundamentally, a system designed to assemble can also disassmble itself. However, nature works in a different pattern and manner. Here actuators, sensors and structural members are compactly packaged in an integrated fashion and protected from the environment. In addition, nature’s compliant materials are capable of large strains without failures. Also, material properties are variable. It is not easy to achieve the complexity and elegance of biological structures but shape deposition manufacturing technique has enabled them to build integrated assemblies with embedded components and material variations.

6.STEPS BEFORE MOVING IN TO THE FIELD OF BIOMIMETICS

(1)   One should have a basic understanding into the mechanisms by which lower animals achieve physical robustness and an ability to accomplish basic tasks such as locomotion despite large changes in the environment. Studies of animal kinematics, dynamics, structural elasticity, muscle activation and sensing provide us an insight to design and control of small robots, sensors and actuators.

(2)   Passive mechanical properties that enable animals to adapt to changes in environment should be employed to the robots so that they can be utilized in multiple environments and multiple tasks.

(3)   These things should be then manufactured and tested before moving on to the software development part. Tests can include locomotion and manipulation of awkward and delicate loads and application can be to get wreckages from ocean floor or say de mining of particular area. 

(4)   New technologies like Shape deposition manufacturing (SDM) should be tried and used to develop structures with complex three-dimensional gravity, embedded sensors and actuators. It should primarily be done to overcome present scenario of using metal components. Individual parts should be taken into consideration before proceeding further.

(5)    Better actuation and sensing signals are the key to better biomimetics. Much research is to be done in that field for improvement of robots.

(6)   Automation of these robots is also another important aspect of this whole process. For that, however, application field is very important without which it is difficult to get an idea of what to develop.

APPLICATIONS

Various exciting applications can be thought of in this field. However, concrete development at this stage is very less. However, I would like to discuss some applications.

-     The most exciting applications of all is that of the use of bioinspired robots in the exploration of Mars. The technology called BEES ( bioinspired engineering of exploration systems )[11] is being planned to utilize in future missions to Mars. Here, bioinspired visual navigation sensors are used on small flyers to enable autonomous flights. The bioinspired sensor suite consists of dragonfly inspired ocelli for flight stabilization and attitude referencing, honeybee-inspired optic flow for terrain following, lateral – drift containment, and localization, and sun and sky polarization based compassing.The sensors also include a perspective camera and panoramic camera for autonomous terrain following and hazard avoidance. Important factors to be considered include processing feature/landmark images and autonomous recognition strategies, sensing and communication, locomotion, adaptive control and reconfiguration, deployment, cooperative operations and bioinspired system design.

-          In military missions, where humans cannot reach, these robots can reach and do the work if possible. For example, an unmanned aerial vehicle can be used for surveying and a lobster robot for checking if there are underwater mines.

-          Further, surveillance capabilities need to be enhanced. Lobster robots can be sent to deep heights under the water to find out unexplored waters. At the same time, it can be used to find wreckages and debris under the water.

CONCLUSION

The field of biomimetics is at a very nascent stage and is yet to be explored to its full potential. However, the important thing that we all understand is that lower level of animals possesses tremendous capacities, which can be slowly brought into the field of robotics. Its not about recreating a cockroach but to recognize its good points so that robots are developed accordingly. Inspiration can be drawn from the fact that some of the top universities of USA – Stanford, UC Berkeley, Harvard and John Hopkins have joined hands to promote this field and work further in the area. Also, if we can combine the best of abilities of all animals on earth the day is not far when we would be able to develop a single robot capable of running faster than cheetah, stronger than the lion, faster than swift in flight and agile as eagle would be developed.

 

 

 

 

REFERENCES

[1] F Ruffier, S Viollet, N Franceschini, “OSCAR and OCTAVE: Two bio-inspired visually guided micro-robots,” Proceedings of International Conference on Advanced Robotics 2003 pp 726-732

[2] Erki Suurjaak, “Bioinspired Hardware”

[3] Jonathan E. Clark, Jorge G. Cham, Sean A. Bailey, Edward M. Froehlich, Pratik K. Nahata, Robert J. Full, Mark R. Cutkosky, “Biomietic Design and Fabrication of a Hexapedal Running Robot”, 2001 IEEE International Conference on Robotics and Automation. 

[4] Robert Full, Timothy Kubow, John Schmitt, Philips Holmes, Daniel Koditschek, “Quantifying Dynamic Stability and Maneuverability in Legged Locomotion”, Integ. And Comp. Biology.42:149-157 (2002).

[5] Robert Michelson, “Entomopter Project”

[6]Ayers., J., Zavracky, P., McGruer, N., Massa, D., Vorus, V., Mukherjee, R., Currie, S. (1998) “A Modular Behavioral-Based Architecture for Biomimetic Autonomous Underwater Robots”.

[7] Biologically Inspired Robotics Laboratory, Case Reserve Western University.

[8] Linda Dailey Paulson, “Biomimetic Robots”, September 2004, Computer Magazine.

[9] “Biomimetic Underwater Robot Program”, Marine Science Centre, North eastern University.

[10] Stanford Centre for Design Research, “Sprawl Robots,” www.cdr.stanford.edu/biomi

[11] Sarita Thakoor, John Michael Morokiaan, Javaan Chahl, Butler Hine, Steve Hornetzer, “ BEES: Exploring Mars with Bioinspired Technologies”, September 2004, Computer Magazine.