Cambridge term has ended and now its easter break. However so, the blog post will continue throughout the break, as long there is substantial progress that is worth posting.
From the last blog post, the main progress has been on writing the TDM and brushing up bits and pieces of the robot. The document is now on google drive, so please have a look at it. Following the TDM upload, links to relevant documents such as CAD files and additional photographs have been uploaded to this website.
During the process of submission of the TDM, the robot was tested for many of its components, in which quite a significant number of problems have been identified, especially with the hardware components of the robot.
For example, with the arm a proof of concept video was taken to demonstrate it opening the tubes specified in the competition (https://drive.google.com/open?id=17-xR0y_-yHTKShBLCCoTHTPmZRMAKFze).
While this test was successful, and all the components worked accordingly, the arm it self was quite soft (as in not rigid), and in return was not able to control with a great deal of accuracy. Although this is partly because of an incomplete control software, there is obvious work to do be done on the hardware.
Therefore, a new arm is being reconstructed now!!
A new, more powerful servo was purchased (ANNIMOS RC Servo High Torque 20KG Digital Full Metal Gear Waterproof Control Angle 180), and all components were newly laser cut/3D printed to accommodate the new servo with different dimensions. This is seeming to be more rigid and “strong”, so it should improve on the controllability of the arm.
Following the redesign of the arm, a new axis will be added. It was realised that with the old arm, it is impossible to look through all the holes on the pipestar, because of a lack of a degree of freedom. Photos and videos on how this new servo will function will be posted on the next blog post as it is being made at the moment.
The other major problem is with the chassis, and the inability to climb over all the obstacles nicely. Yesterday, step fields were created, and was found out the in some occasions (which occur quite frequently), the robot is unable to climb over the obstacle because part of it gets stuck on the bottom side of the chassis, and the wheels are not touching the ground. There are a few ideas around this.
1. Make the wheels bigger and with better grip.
This does not solve this problem completely, but will help once the wheels do touch the ground. From the TDM and other photos, we have used silicone to coat the wheels, but unlike previous methods of just sticking them on with cable ties, we have now a casted version.
It works pretty nice, but not all 4 are made so it hasn’t been tested to its full capacity. It also has a disadvantage of being heavy.
2. Making a tail, or a “push” mechanism explained in the TDM.
The link above is a video, but basically, it is a mechanism to push the robot out of the a stuck situation, which works quite well. However, the downside is that it does require an additional motor, and will be quite bulky on the robot.
This will most definitely be mounted on the back side to assist the robot going over 10cm hurdles, but it would be better if the robot was able to navigate through all other obstacles than hurdles without using it. This is because it will be difficult to mount this on the front due to weight reasons and interference with the arm. However the robot must travel in the reverse direction during the competition, through obstacles as well.
3. Remaking the chassis, but lowering the wheels.
Although this would not be too easy, the chassis can be redesigned in a clever way to lower the wheels from the main base, by lowering it for say 2~3cm with a combination of big wheels, it might be enough for there the be clearance on the bottom of the robot of hitting or getting stuck with any obstacle.
This method will be further investigated through the next week, first by making a simple cardboard shape to figure out where the optimum location of the wheels could be.
A finally change of the hardware of the robot will be to tidy up the insides, because currently it is a mess of wires. One method of this is to make a new board which will accompany many of the components but in a more compact way using a PCB.