Mechanical

Our main focus for this week was to conduct research on the caterpillar track and the components that follow:

The three main components in the track are:

• The gears that transfer power from the motors to the belt
• The belt
• The suspension 

Belt

We found only pre-dimensioned belts online and this means we have to design the model to fit the belt. Custom made belts are out of our budget and are not a viable option.

Alternatively, we are considering a home made belt using duct tape/a long band and cut out pieces of cardboard as reinforcement. This will allow us to work on a prototype for the belt system quicker. Most importantly, we have to design the gears to suit the grooves/holes on the belt. 

Gear

The most important factor here is to avoid any kind of slip between the gears and the belt. If the gear loses its connection we will lose all drive and effectively render the tank powerless. However, what the form of the teeth on the gears will depend on the belt we use. Lastly, we will have motor driven gears on both sides of the belt for evenly distributed power. These motors will be synchronised. 

Suspension

We decided to include some form of suspension for 2 main reasons. Firstly, we wanted to have a more complex system for more of a challenge, and secondly to avoid vibrations as much as possible. The camera we intend to use for recognising our target is not the best and a large amount of vibrations can prevent the system from choosing a target. In addition, a suspension will reduce risk of parts or cables coming loose when driving on uneven terrain. 

We are currently discussing designs for the suspension. The springs and structures will be attached to the lower wheels of the track. 

Computer

From the previous week (week 2)  have we discussed and agreed upon to distribute our own assignment between us four in the computer group. We have already decided  to divide the software and the electrical part in sections and what part we should priorities since some parts are dependent on the other parts. 

Aim and Fire system

Looking into ways to create an autonomous aiming system, the solution we’re currently going for is to program a “room”. This “room” is created by having set limits on how far the barrel can move in any direction by having a predetermined max limit the x and y axis, everything inside these parameters is the “room”. If our camera determines a target within this “room” it will trigger the aiming program adjusting the barrels position and ultimately fire a pellet at the target. How it will achieve this is by signaling the servos connected to an arduino and adjust our servos holding onto the barrel moving it to the correct position where it will hit the target. 


Camera and Object recognition 

After a lot of research we have figured out that we will use openCV on a raspberry pi with a pi-cam for object recognition. This is something we haven’t done before, so we have a challenging path to go. But again, just like those previous weeks we have learned a lot from the researches and discussions.
The second challenging task we think will be with this object recognition is to have a complete communication between these different requirements we have to fulfill. To get the end result we want, we need to get these things to communicate together, which is something we know we will succeed in. Currently, we have just done tests on RC cars and a test chassis made out of wood. These tests has shown different results, which places us in a position where we have to put these results together to get the accurate result we want. We see this process as a type of a long-term plan where we have to do many iterations. 


Autonomous Maneuvering – and Avoidance system 

Since we don’t have the necessary part to build our Tank yet (depends on when machine students are going to complete to build the part of tank). Do we temporary work with available option, like to build an Arduino component on an Arduino car instead of a Tank.  

We have successful got us off with a good start with the Autonomous manoeuvres and the Collision Avoidance system, also known as Collision Mitigation system.  

Unlike simple and easy DIY Arduino car, with limited use. Do we prefer to set up requirements algorithms that will make sure that in the end it will perform completely at ease and effortless, so that our main vehicle will have good driving behaviours and so on.  

Hopefully in the end it will successfully perform well in:

–       Safety measurements

–       Calculated speed control

–       System stability simultaneously

–       The obstacle avoidance path planning approach

–       And hopefully improve significantly in the artificial field

It is still long ahead before we can accomplish our plan.
So in this week did we just improve on our arduino car. Getting the rest of the necessary components and run it on “Arduino IDE”.
After some trials, did we finally get some result and successfully operated the car with just one sonic sonar sensor (temporary, will add more later). With help of the sensor did it detect obstacle and avoided to collide with it.
  Picture 4: Arduino car v.1.