Page 1 - Page 2
Welcome to my latest project - R.A.D.A.R (Reconnaissance And Disaster Ant Robot), a hexapod robot.. This project is my final year University of Liverpool project.
With RADAR I hope to create a hexapod robot which will be a horse and rider robot. The horse will be a BEAM style hexapod, and the rider an OOpic-C microprocessor. The
microprocessor will handle more complex tasks such as IR heat detection at the level of an animal or human, as well as ultra sonic distance ranging. The micrprocessor will also
handle any human intervention to be made remotely to the robot. I am also hoping to mount a wireless camera to the robot.
Today I have been considering the basics of the robot; I need to design the chassis and establish the basic idea for its walking and turning.
The biggest challenge here is to create a walk that improves on the vertical clearance seen in the Queen Ant Hexapod, although I was very pleased with the walking gait achieved with Queen Ant, she did not have good vertical clearance and hence was poor on rough terrain.
The reason for the poor clearance was the use of solid paired legs, if you try to imagine 6 legs all on the ground with a weight upon them, and then you try to lift a leg on one side, its partner leg on the opposite side will experience force driving it into the ground,
impeding vertical clearance.
To overcome this I plan to joint the legs allowing seperate control of each leg's vertical motion. The robot will still use three paired leg sets allowing the horizontal motion for each pair to be controlled with a single motor. In order to then drive the vertical motion
I plan to use linear actuators mounted onto the legs (one for each leg).
The next challenge is how to turn the robot. I am considering two options at the moment, and have consulted the BEAM list for their opinions on this matter.
1. Use the microcontroller that will be on board to turn the robot.
2. Use a second circuit similar to the main circuit, and disable the main circuit during a turn.
I have been sent a link from Harold Ilano (see Harold's site) for another Hexapod walker robot. This is a very interesting
design which uses only 3 motors in total!! It can achieve this because the front leg pair and rear leg pair are connected, this allows for horizontal control of both with a single motor. The
central legs are not given any horizontal motion, they only have vertical motion, this motion provides the vertical motion for the entire robot. Click here for the page
I have also put a little more thought into the turn function. In order to get her to turn I require the horizontal motors to give continuous rotation in one direction, the front and rear legs moving in the same direction, and the center motor in the opposite direction. The linear actuators
controlling the vertical motion would continue their oscillations as before. Therefore I need to disable the horizontal chip oscillations, and run the motors continuously instead.
The robot will turn based upon descisions made by the rider (the micro-controller), and therefore the rider will be disabling the horizontal motion chips, it therefore seems logical to simply use the same process to run the motors continuously.
Hopefully the microprocessor and an IR communications module are on their way to me. I am not sure what size budget I will be given for this robot, so I am having to buy some parts for myself. I want to get started on the build a.s.a.p as currently
there are so many unknowns that I cannot possibly hope to resolve without building it. Today I have been considering the leg materials, and their design. The legs will need to be strong, and possibly sprung to give a second 'knee' which will
allow it to be more flexible with the terrain.
J Wolfgang Goerlich (see his website in the links panel) has e-mailed me after carefully considering my robot turning dilema, he has suggested a Hexacore circuit instead of the one that I have used for Queen Ant. This is new to me so I shall have to study it and see how it goes, but
essentially the Hexacore would allow for greater control of individual motors. Heres the biomorphic map as drawn by JW:
Today I have been looking more into the circuit and mechanics for RADAR. The biomophic map represents in a simple sense the circuit, in order to understand this I recommend reading this article by JWGoerlich.
I have ordered some copper rods from eBay so that I can prototype a leg design for the robot. The leg design is giving me the biggest headache. As mentioned previously I need to overcome the issue with poor vertical clearance seen in Queen Ant, this requires more flexibility in each leg. The options available to me
so far are:
- Seperate each leg, this allows control of seperate legs. Problem: More motors, more weight, more complicated, more power needed!
- Use a 'knee' joint on each leg, the knee joint is controlled by an actuator. This would require an actuator on each leg, but allows each leg pair to remain controlled by a single horizontal motor.
- Use a simliar design to Queen Ant, three segments to the body, each holding a leg pair, two motors then control vertical motion by twisting the segments. To improve on this I propose using jointed legs
held with a spring. This (in theory) would have enough give in it to allow the segments of the body to twist further giving better horizontal clearance. The difficulty here is getting springs of the correct tension, too tight
will not allow any movement, and too lose will cause the robot to collapse!
The copper rods have arrived which will enable me to begin prototyping the leg design. I have also ordered some springs which should arrive soon. I have been putting a lot of thought into the leg design, it is very difficult when I cannot model it
or look at it physically. Here is the design that I shall be prototyping:
Today the springs arrived, however I am not sure if any of them will be up to the task! It seems to be very difficult to find a good supply of springs. Last night I started the leg prototype, heres the pics (click for fullscreen):
The part I have created is the foot section, this will now need a pivot and a spring attaching to it.
Continued - Page 2 of Hexapod human seeking robot