This project started off like many the desire to build and the curiosity to see if it would work. Thanks to both the desire to build and the curiosity the project came to life. Oh and I almost forgot thanks to DFRobot for the Devastator base. The object of this was to fit a netbook or tablet on this small base. The base itself is well constructed and very sturdy. The main portion of the base is made out of metal. The kit comes with all the hardware, drive motors, tank tread, toggle switch, female power jack and tools to boot. The current base is adjustable and able to house a 8" tablet to a 10" notebook. The battery pack is a 4.8v 3000 mah Nicad battery pack. The servo control board is from DFrobot, its the Veyron 24 channel. How do we drive the geared motors. I used two Tower Pro 9g s90 servos. After taking the two servos apart I discarded the motor in each. Next I measured the potentiometer to see what its resistance was and replace it with two resistors. I then soldered new wires for the output to the motor and cut a small slit in the servo case to pass the wires through. What we end up with is a nice little motor controller in a case. The software that it is currently running on is designed in .NET and is my first try and making a piece of software easy to use. While it still needs some refinement it does work. The goal on the software is to make something that a child could use to start learning how a robot works. Making something easy to use is not so easy. The user presented the whole process in steps. Step 1: we select a delay start time, the servo channels to use and the comm port. Upon completion of the first step the second step appears. Step 2: Select the direction of travel, select the speed of travel and then the time to travel. Steps 3, 4, 5 and 6 are all repeats of step 2. after each step the user is given the ability to save what they have just done or reset it or run it. You can save it and then run it and then load it back in and modify it some more until you get the desired results. While this is no great achievement in robotics, it is an interesting project and has presented me with even more curiosity.
UPDATE: Recently I have been using the software I designed that detects and follows colors. For my initial test I used a green piece of poster board and placed it in an area away from the view of the webcam. The RoboTank roams around until it detects that color and then goes directly to it. The next test was to move the poster board and see if the TankBot could follow it. Following does occur but I need to change the reaction time the software has. The software tends to react to the movement of the poster board quicker than I can move so it sort of over shoots the movement of the poster board and then has to correct its self which makes the RoboTank move left, right, left, right instead of doing slight right or left and moving forward. I believe I can correct this issue by changing the speed of the motors so the reaction time it slightly slower. Once this is working correctly I will then move onto using my object detection software and will post results of this.
UPDATE: I have moved the Veyron servo controller to the outside of the unit. Reason for this is that it makes more sense to have it on the outside so that I am able to get to the servo connection and also I plan on using the Veyron on another project and I didn't want to have to take apart the tank chasis to get to the board when the time came. This will also make it easier to add the two addidtional servos I plan on adding for the control of the cam. I have also added a angled bracket made out of PVC to hold the Veyron and two screws with rubber sleeves and washers to retain the USB cable.
UPDATE: I have finished designing the software that will now run the RoboTank. The software does color detection. So you select a color and it will follow it. Within the software are features to adjust video frame rate (FPS), Servo speed and Threshold and Performance of pixel reading. The video screen is divided into sections so the further left the color is in the screen the sharper of a left turn it makes to get to the color and the same applies to right. When there is no color detected it stops until the color returns.
Explore the new Raspberry Pi RP2350 microcontroller, featuring Arm Cortex-M33 and RISC-V Hazard3 processors, expanded memory, and enhanced security features. Perfect for advanced embedded systems.
Learn how to run YOLOv10 on the Unihiker Board with this detailed guide. Optimize object detection for real-time applications in embedded systems with ONNX conversion.