Computer  

This week we are going to test the manual control system of the tank. The purpose of this test, is to understand the behavior of the tank and to make sure that the manual control system works along with the components. Since, this is a tank and not a car, we also need to make sure that the motors on the left side works simultaneously, and the same on the right side. If these motors are not working simultaneously, it may damage the belt.

The first step in implementing the aim and fire system was manual control of the servos. Building upon last weeks idea, both of the servos were connected to a joystick. As the servos can only operate in one plane of motion using one joystick for two servos work perfectly, as the arduino joystick has its own connectors for the x and y axis. This made managing the signals sent to each servo a non-problem as long as the wiring didn’t get messed up. 

For the object recognition part, we have few problems with downloading the software. This software has to be downloaded directly to the raspberry pi by the internal terminal, where we have to use the right commands to start downloading packages, libraries, software and so on. One of the problem is that we are trying to download the newest version of openCV, which is version 4.1.1, but there are some old packages we need to substitute with new packages. There has been many iterations with the object recognition to get the software to work, and we are still on it. We also tried version 3.4.0, but somehow that wouldn’t either work. So, now we are testing pi zero instead of pi desktop, to see if that can help. As said, there are many iterations to do to get to the right product and find the right solution. 

Goal for week 4 is to continue to work on:

1. Camera detection, Object Detection, Lock target 
2. Aim and Fire system 
3. Autonomous Maneuvering – and Avoidance syste

Mechanical

As composite materials are part of our syllabus this semester, we would prefer to make the  body of the tank out of a composite material, either carbon fiber or glass fiber depending on the cost. We began the week by further reading into how to use composites for complex forms and with some discussion with Kåre, we agreed that it should be possible to make the body in one casting. However, to do this there are a few considerations we have to take when designing further.  

These are:

• To cast a model in composite we require another model to cast the laminates on. This means we have to either laser cut/3D print the model in parts and assemble it to our specifications. 
• Another difference when comparing to laser cutting is we have to avoid sharp edges and corners as much as possible as casting around these points are difficult. 
• This can be a time consuming process, so getting the physical reference model ready as soon as possible is a priority. 

Secondly, this week we have finished modelling a possible suspension system and we must now acquire springs with enough stiffness to support the weight of the tank with and without stress. To do this, we plan on getting a variety of springs and testing with the 3D printed model of the suspension. We are encountering some troubles regarding the shooting aspect of our project. We are currently in the planning phase, and we plan to use an electric motor driven airsoft gun. Our idea is to disassemble the gun and reassemble it within our design to avoid having an actual pistol in the tank. As for our belt, we have decided to make it out of duct tape and cardboard, inspired by a youtube video. However, we cannot proceed with precise modelling as we need the dimension of the gun as soon as possible. Purchasing it however, will have to come after discussing our budget.