Robo-Sloth: A Rope-Climbing Robot

Size: px

Start display at page:

Download "Robo-Sloth: A Rope-Climbing Robot"

Maurice Newton
5 years ago
Views:

1 Robo-Sloth: A Rope-Climbing Robot Sandeep Urankar, Pranjal Jain, Anurag Singh, Anupam Saxena and Bhaskar Dasgupta Department of Mechanical Engineering Indian Institute of Technology Kanpur Abstract The design of a rope climbing robot was an exercise in bio-mimicry, an attempt in copying the exact motion of a sloth bear while climbing. The advantages of bio-mimicry are well known, but the true justification is that since much of the proposed robotic behavior already has precise analogies in nature, it is reasonable to copy them for our applications. Biological systems, having evolved over millions of years, provide us proven and efficient solutions to problems of navigation, motion, sensing and searching. Such a design based on a biological organism, if perfected, would be much better than its conventional counterparts. The design which results from this exercise is quite robust, versatile and has many potential applications in industry. 1. Introduction A study of the sloth bear (Fig 1) in motion reveals that the bear uses both pairs of its limbs (fore limbs and hind limbs) to climb a rope or a tree. It moves both its hind limbs in one step and then both its fore limbs in another. Each pair of limbs acts as a gripper. Only one pair of limbs is used at a time for gripping the tree (while in motion) while the other pair slides over the tree. The bear uses its back as a hinge, the back being straight in one instant and bent in the other. This movement when coordinated can be a very efficient and a reliable mechanism for climbing. No wonder then, the sloth is one of the best climbers in nature. This is also the reason why it was chosen as a model for our study. In recent years, several researchers and groups have reported work on such biomimetic devices. For example, The Robo-tuna (Barrett [1]) is an ongoing project at MIT s department of ocean engineering, where attempts are being made to copy the motion of a blue-fin tuna in order to develop a better propulsion system for autonomous underwater vehicles. Ayers et al [2] reported attempts being made at the Marine Science Center, Northeastern University, to copy the behavior of lobsters for the purpose of conducting autonomous investigation of both the bottom and water column of the littoral zone of an ocean. Analysis and development of snake-like devices has been reported by Hirose [3]. The ROBO-Sloth is comparatively simpler as far as its behavior is concerned. It has only to move along a pre-defined path. An improvement in design would be to enable the robot to shift between ropes, whereby the problem would become that of motion in space rather than along predefined path (rope). 1

2 Fig. 1 The Sloth 2. Mechanical Model 2.1 Design of Backbone A mechanical model, formulated based on the observation of the sloth, would essentially be a set of four bar mechanisms (see Ghosh and Mallik [4]). The entire body of the bear and the grippers can be modeled as four bar mechanisms, performing actions similar to that of the bear. Beginning with the backbone, it can be modeled if it were to be thought of as two links hinged at the center. This is of course an approximation, since the bear s back is more flexible than this. However, this approximation is valid and serves our purpose at the moment. In future, if needed, more redundancies may be added. Fig. 2 Backbone of mechanism The links would have to rotate about the central joint in order to mimic the sloth. Not only that, we would need to control this rotation if we are to control the climbing. For this purpose, a power source has to be placed at the joint (Fig 3). A motor could be placed between the two links. In the actual fabrication of our robot, we chose to place a worm and worm-wheel mechanism at the center. The reasons were obvious; this is a self locking mechanism, it ensures that the link doesn t fall back into place once the motor has been deactivated. Besides this obvious advantage, it can provide the appropriate speed reduction in order to make the motion of the sloth slow and steady. Choosing the appropriate power transmission mechanism at the centre is crucial as a lot of torque acts at this point, especially if the robot is to carry huge payloads. motor 2

3 Fig. 3 Links with motor in place Two more pin joints and links are added to the ends of the backbone, (Fig. 4). The new links are now constrained to move along a straight line or a fixed path which approximates the rope to be climbed (Frame). At any given instant, link 4 is fixed while link 5 slides along the fixed path or vice versa. The whole mechanism now is a four bar linkage, powered at the center. Links 4 and 5 are important as they act as sliders, whereas links 2 and 3 act as cranks. Links 4 and 5 also provide the base on which the entire gripping mechanism is to be mounted. Even if the backbone, without links 4 and 5, is complete, they are added to the backbone for reasons detailed above. Link 4 Link 5 Link 1 Link 2 Link 3 Fig. 4 Four bar mechanism (backbone) 2.2 Design of Grippers The design of the gripper is an essential part of the rope climbing robot. During the vertical climb, the gripper has to bear the entire weight of the robot and the payload. This, is the main consideration for the design of the gripper. The gripping mechanism can be approximated by another four-bar mechanism. This mechanism is to be attached to links 4 and 5 of the backbone. Link 1 Link 4 Link 2 Link 3 Fig. 5 Gripping mechanism The four links, shown in fig. 5, are joined together by pin joints. The mechanism can be powered by means of two pulleys. A motor is attached to the smaller pulley and through a transmission belt the rotation is conveyed to a bigger pulley. The larger pulley is attached to the four bar mechanism, when it rotates, it causes link 3 to clamp or unclamp the rope held against the base. 3

4 Pulley 2 (large) a Link 3 (clamping link) Pulley 1 small Fig. 6 Parameters a, m, l The parameters l, m and a are important (Fig 6) as they determine the level of frictional force finally generated at the clamp. If the length of l is longer than that of m, more leverage will be achieved at the point of contact. The angle a will determine if the gripper is self-locking or not. The gripper has to be designed such that, it is self-locking. This ensures that the robot does not slip at any point of time. More over this property allows us to put large amount of payloads on the robot. The more the payload, stronger will be the grip. The figures 7 and 8 illustrate how the gripping mechanism becomes self locking. As soon as the robot begins to slip in the direction indicated, the reaction force (friction, also indicated) will act in a direction so as to oppose the direction of slip. This force generates a torque about the pin joint such that it causes the clamping link to rotate in such a direction that the link clamps the rope under the action of its own weight. Hence every time the robot begins to slip, the clamp will grip it tighter, making the mechanism self-locking. F= Reaction force T= Torque F T mg Direction of slip of robot Fig. 7 Self locking gripper (during horizontal climbing when weight acts downwards) 4

5 F T mg F= Reaction force T= Torque Direction of slip of robot Fig. 8 Self locking gripper (during vertical climbing) 2.3 Design of guides Guides are needed in the design so that the entire weight of the robot does not fall on the clamping mechanism. The normal force from the rope will produce a torque such that it will unclamp the gripper. In figure 9, shown below, two rollers are placed on either side of the clamp such that when the robot is suspended from a horizontal rope, the robot hangs on the rollers. The rollers will allow the rope to pass freely between the clamp (rolling friction is minimal and can be ignored) Guides 2.4 Complete Design Fig. 9 Guides The design of the backbone and the gripper completes the mechanical design. Figure 10, below, shows the complete design with the gripper bolted on to the backbone. It also shows how the three motors are bolted on to the backbone and the assembly of the worm and wormwheel arrangement. A gear box (shown in figure 10) is required at the center because the worm has the tendency to dis-engage from the worm-wheel under large loads. 5

2.5 Design of control system Fig.10 Final assembly As opposed to many other bio-mimetic systems, the design of the control system for the Robo-sloth is fairly simple.

6 2.5 Design of control system Fig.10 Final assembly As opposed to many other bio-mimetic systems, the design of the control system for the Robo-sloth is fairly simple. It has only to move back and forth on a well defined path, a rope or a wire. However if it were to be designed such that the robot could shift between two wires close to each other, the problem of controlling would become all the more difficult. For the present design, figure 11 shows the algorithm for motion. There are many ways to implement this algorithm. An electronic circuit was designed to control the Robo-sloth. However, the same task can be achieved by an RCX available in the Lego mind-storm kit. The programming is iconic and simple, allowing the designer to make changes at will. Here we a Start side motor1 and motor 2 (hold upper gripper and lower gripper) Halt Direction of motion Up down Release lower gripper Activate central motor in clock wise direction 6 for T seconds Release upper gripper Activate central motor in anti clock wise direction for T seconds Hold upper gripper

3. Fabrication and Assembly Step 1, the worm and

with the gearbox in between. (Fig. 14, 15) Fig.

15 Step 3, two kinds of motors are used in the robot,

The central motor is passed through two rings which

through channels with key slots cut in them, the

respectively. (Fig. 16, 17) Fig. 16 Fig.

7 3. Fabrication and Assembly Step 1, the worm and worm-wheel are assembled into the gearbox (Fig 12, 13) Fig. 12 Fig. 13 Step 2, four links of the backbone are assembles with the gearbox in between. (Fig. 14, 15) Fig. 14 Fig. 15 Step 3, two kinds of motors are used in the robot, the motors have to be bolted to the backbone. The central motor is passed through two rings which are bolted to link 2, and the side motors are passed through channels with key slots cut in them, the channels are then bolted to links 4 and 5 respectively. (Fig. 16, 17) Fig. 16 Fig. 17 Step 4, grippers have to be assembled using shafts and bushes. The grippers have to be bolted on to links 4 and 5 respectively. (Fig. 18 and 19) Fig. 18 Fig. 19 (Duplication of Fig. 10) 7

4. Results Based on the principles and designs discussed in the previous section, the Robo-sloth was fabricated. The parameters of the robot and the steps of motion are discussed in this section.

8 4. Results Based on the principles and designs discussed in the previous section, the Robo-sloth was fabricated. The parameters of the robot and the steps of motion are discussed in this section. The material chosen for its backbone and gripper was aluminium. Aluminium was chosen because it is much lighter than other conventional metals available to us. Brass had to be used when there was a mating between two parts. E.g. worm and worm-wheel, shaft and bush. The weight of the sloth was measured to be 1.7 kg and a payload of 5 kg was carried successfully. The payload is restricted by the huge torques generated at the centre. An improvement in the central joint and a stronger central motor would lead to an improvement in the payload capabilities. The power required for normal operation is 20 W and the sloth can run on a 12 V DC supply. This indicates a 12 V battery can be carried on board, but the weight of the battery would have to be deducted from the payload weight. The current design, however, did not carry the power supply on board. The operational speed capacity of the sloth is 2.7 m/min. The comparatively slow speed of operation was chosen intentionally to mimic the sloth. The tip to tip length of the sloth is 62 cm, its breadth is 7 cm and the maximum height it can achieve is 14 cm. The sloth is programmed to operate between backbone angles of 60 deg and 150 deg. Having discussed the parameters, the steps of motion are discussed further. The photographs shown below, are slides from a movie that was made of the Robo-Sloth in motion. The slides show, how successful the Robo-sloth has really been in mimicking the real sloth. It was tested many times and the performance was found to be highly repeatable. Note the angles of the gripper links, they indicate opening and closing of the grippers. The bending of the backbone indicates the motion of the robot. Step 1, both grippers are locked and backbone stationary (central motor switched off). Note the angle of the backbone, (almost 60 deg), climbing begins at this angle. Also note the positions of the clamps; they are both in gripped position. In this state, the robot is ready and waiting for a command to move. Clamp 1 Side motor 1 Central motor Fig. 20 Picture of Robo-sloth 8

Step 2, when Side-motor 1 is switched off, the respective gripping mechanism

21, When the Central motor is switched on, link 5 begins to slide to the right.

22 Step 3, when the extreme position is achieved, as in fig 22, side-motor 1 is

The gripper on the left unclamps (fig 23).

This continues till the backbone reaches its other extreme position (almost 40

9 Step 2, when Side-motor 1 is switched off, the respective gripping mechanism unclamps. Note the position of the clamp in fig. 21, When the Central motor is switched on, link 5 begins to slide to the right. The link continues sliding till the backbone straightens out, note back- bone angle in Fig. 22 (Almost 180 deg). At this point, the central motor is switched off. Fig. 21 Fig. 22 Step 3, when the extreme position is achieved, as in fig 22, side-motor 1 is switched on and side-side-motor 2 switched off. The gripper on the left unclamps (fig 23). Central motor is switched on, Link 4 begins to slides to right (fig 24). This continues till the backbone reaches its other extreme position (almost 40 deg) (figure 24).Next, the robot returns to its original state and the entire cycle begins again. Fig. 23 Fig. 24 A similar performance is achieved when the robot is climbing vertical ropes (see Fig 25, 26,27). Fig. 25 Starting position Fig. 26 Extreme position Fig. 27 In transition 9

10 5. Conclusion The sloth bear s motion was studied in detail and a kinematic model was made. The kinematics, after suitable approximations, boils down to two four-bar mechanisms: the backbone and the gripper. The linkages were made self locking so that the robot would not slip under its own weight. A control system was designed such that the gripper and the backbone when operated in coordination would simulate the climbing motion of the sloth. Every attempt was made to reduce the weight of the robot so as to maximize the payload capacity. The speed of operation was intentionally kept low so that while in operation the Robo-sloth could be slow, silent and inconspicuous. The self locking nature makes it very reliable. These properties make the Robo-sloth an excellent spy. Preferably under suitable camouflaging packaging, it could be used very effectively for monitoring buildings. It already has a good payload capacity; therefore it may carry the power supply and a camera onboard. If the gripper were to be replaced by suction pads, it would have the capability of scaling walls as well. The design could be slightly modified to carry grippers and suction pads at the same time. Robo-sloth would then become a really versatile climber. It would have the distinct advantage of being able to maneuver around corners, being able to crawl on the ground, wall and the ceiling and shift to wires, pipes and cables when required. It could also crawl through AC ducts very easily. Hostage situations in large buildings could use such a monitoring device. Maintenance of high tension power lines could be another application. Maintenance of lift shafts and cables could also be done through such climbers. To conclude, biology has proven through evolution that the mechanism of the limb is far more versatile than that of the wheel. The task of climbing a rope could have been performed efficiently with a wheel mechanism but the potential of the Robo-sloth or any other biomimetic device is much more. If the hinge joints in the back bone of the Robo-sloth were replaced by ball-and-socket joints, then the robot would have much more flexibility, to shift between ropes and turning would become possible. There is no limit to the range of maneuvers it may then perform. 6. References 1. David Barrett. (1995) MIT ocean Engineering Testing Tank, Biomimetics Project: Robo Tuna, 2. Ayers, J., Witting, J., McGruer, N., Olcott, C., Massa, D. (2000) Lobster Robots. In: Proceedings of the International Symposium on Aqua Biomechanisms. T. Wu and N, Kato, [eds], Tokai University. 3. Shigeo Hirose. (1993) Biologically Inspired Robots (Snake-like Locomotor and Manipulator), Oxford University Press. 4. Ghosh and Mallik, Theory of mechanisms and machines 1998, Affiliated East-West Press Private Limited. 10

Model No: TC10. Parts Information: Tyre Changer - Automatic

Model No: TC10. Parts Information: Tyre Changer - Automatic Page 1 of 11 1 TC10.01 BODY 2 TC10.02 COLUMN 3 TC10.03 HORIZONTAL ARM ASS'Y 4 TC10.04 WASHER 5 TC10.05 RUBBER FOOT 6 TC10.06 COVER 7 TC10.07 SCREW M14x42 8 TC10.08 PRESS COVER 9 TC10.09 STOP-UP 10 TC10.10