Rubin Cholera Under the guidance of: Prof. Justin Seipel ME-597 Bio-Inspired Robotics ASSIGNMENT No.

1
Topic of interest Bio-Inspired stored energy legged robots Note: There are two kinds of research papers studies here. The first few are studies of some of the best jumping animals, and the rest are research papers on biologically inspired jumping robots. Some findings that I found really interesting have been highlighted. 1. Leg design and jumping technique for humans, other vertebrates and insects, by R. McN. Alexander This research paper makes two models of vertical jumping (a two jointed model and a three jointed model) and compares the three jumping techniques (squat, countermovement, and catapult) by varying parameters like leg length, muscle properties, and stating angles. Some findings that interest us are: For jump height catapult>countermovement>squat. Three jointed leg model shows a significant advantage in jump height over two jointed legs. Increasing leg length may not necessarily lead to increase in jump height (flea vs locust) Humans exert 2-3 times their body weight on ground while jumping and frogs exert only 3.5 times, but the jump performance is vastly different owing to differences in technique, muscles, and leg lengths, and starting angles (whereas fleas exert more than 100 times their body weight) There are many other findings pertaining to muscle properties which are of no interest to us.

Some of the gaps/future work are: They consider only a vertical jumping model for jumping. They consider equal leg lengths above and below the knee. They applied many biological and physiological limitations to the jumping model which we need not apply to make an optimized jumping model for robotics.

2. Froghopper: insects leap to new heights, by Malcolm Burrows This research paper studies the dynamics and technique of a froghoppers jump. Supported by measured values of forces, power, acceleration etc it compares the froghoppers jump with other high performance jumpers like fleas and locusts. Some interesting statements are: Jump performance is increased by either by having long legs to increase the levering power (frogs, bush babies), or by releasing stored energy (flea), or both (grasshopper, locust). The hind legs have a unique locking mechanism to lock the legs just before the jump so that the stored energy is released rapidly. But due to this they not used for normal locomotion and merely dragged along during normal walking. Force exerted is 414 times their body weight

Some of the gaps are: Mechanical or physical modeling was not done.

3. The energetics of the jump of the locust Schistocerca Gregaria, by H.C. Bennet-Clark This research paper is a complete study of the jump of a locust. He recorded the jumps of a sample lot of locusts and stimulating them to jump on a marked surface. He also studies the bio-mechanics of a locust by experimentally finding the mechanical properties of the muscles and the energy storing member. Then the parameters like velocity, energy, force were calculated using the data gathered from the experiment and various graphs were plotted to see the effect of some varying the parameters. Some interesting statements are: The curve of energy released vs the femur length was plotted showing a roughly increasing curve. At low temperatures jumping is not as affected as walking due to stored energy.

Some gaps/future work are: Take off angle of 45 degrees is assumed for all calculations in the model. The electrical stimulation used does not necessarily reflect a natural jump by a locust. Physical imitation models were not made to validate the findings.

4. The jump of a flea: a study of the energetics and model of the mechanism, by H.C. BennetClark This research paper decoded the age old mystery of the jumping performance of the flea with the help of filmography. This used the same setup described earlier of electrically stimulating the flea into jumping. Some interesting points: The flea slowly stores the energy of hind leg muscles in a springy resilin element. Then it locks the energy for 0.1ms and releases it as an impulse The landing mechanism for the flea is the same regardless of whether it is a vertical jump or a long jump, or if it hits a wall. It lands by extending its middle and hind legs sidewards and the front leg downwards The flea is able to direct its jump by varying the angle of hind legs.

Some Gaps/future work are: It focuses

5. A mechanical analysis of the hind leg of a frog (Rana temporaria), by LESLEY J . CALOWA ND R. McN. ALEXANDER This research paper extensively studies the hind leg of the robot for jumping and swimming. It delves more into solving the mechanics of how the muscles operate. The properties of the muscles are calculated by analyzing the data from filmography, the ground reactions, and stress, strain tests on the muscles. Some interesting findings are: The position and posture of the frog obtained from filmography when compared with the ground reaction force readings with respect to time, shows that maximum force is already transmitted while the final 2 links are still on the ground. This suggests that the final link of the 5-bar frog leg does not greatly contribute to the jump performance.

Some Gaps/future work: The focus was on the biomechanics of frog hind legs. Like one of its major findings was A theory is presented which indicates that if a typical vertebrate muscle is used to

accelerate a mass from rest over a given distance, the final velocity of the mass will be greatest if the moment arm of the muscle is such as to make the stress, at the end of the period of acceleration, about 0.2 of the maximum isometric stress. This finding is irrelevant for us, as we are not bound by such limitations. Our limitations are different depending on the type of stored energy we decide to use.

6. A Small, Insect-Inspired Robot that Runs and Jumps, by Bram G. A. Lambrecht, Andrew D. Horchler, and Roger D. Quinn (Biologically Inspired Robotics Laboratory, Case Western university) This research paper is about converting their popular mini-whegs robot to give it an additional jumping ability. It is achieved my adding an independent 4 bar jumping mechanism. This mechanism is powered by an extra motor and transmission box. They also made many changes to improve the running ability of the robot as compared to previous models. Some interesting points are: For a robot that is 10 cm tall and weighing 191g it was able to jump 15-18 cm. The robot is a successful integration of a running and jumping robot, which combines the advantages of both. The insect inspired running model is steerable.

Some gaps/future work are: The 4-bar mechanism had only one degree of freedom (so that a single actuator can control it). The jumping mechanism contributed to 44% of the weight! (it claimed to be inspired by the froghopper whose jumping legs are ~18% by weight and exerts 414 times the body weight times force on the ground). The take-off, flight, and landing cannot be varied. Optimisation of the links was not carried out. The 4-bar had equal length links.

7. Design of a Miniature Integrated Multi-Modal Jumping and Gliding Robot, by Matthew A. Woodward and Metin Sitti (CMU)

This research paper is about making miniature a robot that uses two 4-bars for jumping and it uses the same 4-bars for gliding too, making it multi modal. It does so by attaching a wing-like membrane to the 4-bars, then the 4-bars rotate from vertical to horizontal position during flight. Some interesting points are: They did a system optimization with respect to springs, i.e. no. of springs vs. wire dia vs mean dia. The joint design was inspired by the human knee for weight reduction and high twisting resistance and ability to support routing wires through it.

Some Gaps/future work are: Only Vertical Jumping No controls (for takeoff, gliding and landing) Though the abstract claims that the robot can jump 6m, later in the article they specify that they only tested the jumping + gliding on a low performance model which could jump only 1.7m and the 6m was achieved in the robot without the additional weight of the gliding equipment. Why they did so was also vague. Optimisation of the links was not carried out. The 4-bar had equal length links. Since it is miniature, carrying payload will severely affect performance.

8. Development of a Controllable and Continuous Jumping Robot, by Jianguo Zhao, Ning Xi, Bingtuan Gao, Matt W. Mutka, and Li Xiao (MSU) This research paper also focuses on making a robot that can jump repeatedly (can be prepared to jump again after landing), and also be steerable to control direction of jump. The jumping mechanism consists of two 4-bar mechanisms powered by torsion springs. The shape of the robot is such that when the robot lands it will always fall on one of its 2 larger sides. The mechanism to right itself after landing consists of 2 motor driven links that push on the ground. When on its side it is able to move and steer with the help of a gear that acts as a wheel. Some interesting points are: One of the few 4-bar based robots where the links are not all of the same length. The steering function is multi-modal. The same gear that is used to charge the springs is used to move the robot when it is lying on its side. The gear is placed on one side so when it rotates it acts as a uni-directional steering mechanism.

At 20g in weight with a size of 6cmx8cmx2cm it could jump upto 55 cm. Also it could perfoem 285 jumps on a full battery!

Some gaps/future work are: The parameters of the 4-bar jumping mechanism were not optimized. The takeoff angle is fixed at 75 degrees. There is no control during flight, even though it partially makes up for it being steerable Since it is miniature, carrying payload will severely affect performance. In situaltions with an obstruction on one side of the robot will negate the self-righting mechanism

9. A miniature 7g jumping robot, by Mirko Kovac, Martin Fuchs, Andre Guignard, JeanChristophe Zufferey, Dario Floreano (EPFL) This research paper focuses jumping robot with the best jumping performance. The jumping mechanism is a 3-bar mechanism powered by a torsion spring. The energy storing and release mechanism is an innovative cam based mechanism that charges the spring and releases the energy in one rotation of a cam. Some interesting points are: For a 7g, 5cm height robot it could jump 1.38m (27 times body height) which is the highest per size. The main link of the 3-bar mechanism was optimized. With a payload of 3 g it was still able to jump 1m in height (by changing the initial deflection of the spring)

Some gaps/future work are: The takeoff angle is not adjustable It is not self-righting after landing It does not have control during flight

10. A Miniature Jumping Robot with Self-Recovery Capabilities, Mirko Kovac, Manuel Schlegel, Jean-Christophe Zufferey and Dario Floreano (EPFL) This research paper takes the above mentioned high performance jumping mechanism to give it self-recovering ability so that it can be used for continuous jumping.. The achieve this by

enclosing the jumping mechanism in a spherical steel cage to give it the ability to roll and selfright. Some interesting points are: The self-recovering mechanism works by having the center of mass offset from the center of the sphere, and as the robot charges for the next jump, the legs are pulled closer to the center of mass causing self-righting. Even though the total mass is more than doubled by the addition of the self-recovering mechanism it was still able to jump to a maximum height of 76 cm. Four self-recovering mechanisms were evaluated by using weighted comparative evaluation to choose the best one.

Some gaps/future work are: The gaps are the same as above except for the self-recovering system.

11. The EPFL jumpglider: A hybrid jumping and gliding robot with rigid or folding wings, by Mirko Kovac , Wassim-Hraiz , Oriol Fauria , Jean-Christophe Zufferey , Dario Floreano This research paper takes the same robot used in 9 above, and gives it gliding ability to control the flight. They achieve it by integrating it with multiple wing designs (fixed wing and folding wing) and compare the performances. Some interesting points are: They did a weighted evaluation to compare the various folding wings and fixed wing designs. The fixed wing design was more stable, increased their range and was lighter. This robot does not need a self-righting mechanism because the wings keep it upright while landing.

Some gaps/future work are: Tail and wing control could be included to steer in flight Fixed take off angle. The other gaps are same as mentioned in 9

12. Crawling and Jumping of Deformable Soft Robot, by Yuuta Sugiyama and Shinichi Hirai (Dept. Robotics, Ritsumeikan Univ.)

This research paper tackles the problem of overcoming obstacles with a unique deformable body robot that can crawl and jump. Deformable robots have a distinct advantage over rigid body robots of being able to overcome unpredictable rough terrain. It works on the principle of deforming the robot from a stable spherical shape to an unstable shape changing the center of mass and causing motion or causing it to jump if the deformation is raises the elastic potential energy. Some interesting points are: They evaluate the different types of deformation that causes jump and evaluate the performance. They were able to jump upto 3 times the diameter and climb a slope of 20 degree inclination

Some gaps/future work: The robot worked on an external power source. There are many different materials and deformation geometries that could be tried to improve jumping and crawling performance. Maybe a combination of rigid and deformable body design could be tried.

13. A Bioinspired Concept for High Efficiency Locomotion in Micro Robots: the Jumping Robot Grillo, by Umberto Scarfogliero, Cesare Stefanini, Paolo Dario. The Grillo is a piston based jumping robot that is powered by springs. It has a motor that charges the spring and a trip lever that releases the piston once the springs are charged. Some interesting points: It has elastic front legs to help absorb the shock of landing. They have done a detailed mechanical modeling of the grillo including the stiffness and damping coefficients and rotational stiffness of all the legs for studying the impact of landing.

Some gaps/future work: It does not have the ability to steer It does not have the ability to control takeoff It does not have control on flight

14. An Asymmetric Robotic Catapult based on the Closed Elastica for Jumping Robot, by Atsushi Yamada, Masamitsu Watari, Hiromi Mochiyama, and Hideo Fujimoto The closed elastica robot is an innovative jumping robot that uses an elastic band for powering the jump. The elastic band is attached to the chassis at one end and a motor at the other end. The elastic band is longer than the chassis so the elastic band is at steady state on one side of the chassis. Then the motor rotates one side of the band and after a dead center the band snaps and applies force on the ground. Some interesting points are: It used an asymmetric elastic band to achieve forward motion with jumping. For a 30g robot they achieved 20 cm jump height and 70 cm jump length

Some gaps/future work: It was externally powered and controlled. The battery was heavier the the entire robot No control on takeoff No control during flight It does not have the ability to steer.

15. Jumping robots: a biomimetic solution to locomotion across rough terrain, Rhodri Armour, Keith Paskins, Adrian Bowyer, Julian Vincent and William Megill (university of Bath) This research paper discusses two different jumping robots developed by this group. One is the Jollbot which uses semi-circular metal hoops to make a ball like structure with the battery, charging, and stabilizing mechanism inside. It works by deforming the spherical ball into an ellipsoidal shape and releasing it to jump. It jumps along the axis of the ball and the direction is controlled by controlling the center of mass. The other robot is the Glumper which is a jumping/gliding robot. The jump is powered by a simple 4 bar mechanism and gliding is enabled by covering the 4 bars with a wing like membrane. The interesting points: The jollbot can roll by continuously shifting its center of mass which is also used for directing the jump. The jollbot uses the spherical steel hoops as a mode of self-recovery and as compliant members for jumping A detailed experimental analysis was carried out on the glumper to see the effect of payload and wings vs no wings

The gaps/future work: The jollbot cannot be controlled in flight and the glumper cannot be steered for giving takeoff direction.