Simple IoT Autopilot robot.
Team members: Jonathan Lundström, Alex Norrgård and Oliver Olin
Our goal is to build a car with wireless steering and which is controlled by smartphone app, with button to turn on a simple autopilot (Works with radar = Servo + Ultrasound sensor).
Parts needed int the project + Some jumper cables and a battery of choice (5v-12v), Project used 4 x AA batteries
Circuit diagram of the robot
Both the H-bridge and the arduino board are powered by 4x AA 1,5 V batteries. NOTE: RX VOLTAGE ON BLUETOOTH MODULE PIN 3,3 V, VOLTAGE DIVIDER NEEDED
The process to build the robot is quite straight forward. First the shell of the robot must be 3D-printed, it will house a major part of the electronics. Meanwhile the shell is being printed it is good to check all the electronic parts one by one, so you know everything is working. Then it is time for some soldering and connecting everything according to the circuit diagram. Jumper cables are a good alternative at least for testing everything and can be used for the final product, but more permanent connections could be done. After everything is connected it is again good to test that everything works. As a last thing the code and the app must be done, either by downloading ours or making them by yourself. Our code for the Arduino contains all the needed functions ready for use. Same goes with the app, it is developed using MIT App Inventor and the required code can be seen in the pictures.
STEP 1: 3D-printing the shell
STEP 2: Putting together the electronics
STEP 3: Compiling Arduino code, and building the app
Available for download, in "CODE", section or viewable by clicking "expand suorce".
NOTE: Group found use of the ServoEasing.h for smooth Servo movements between angels. (Also helped with servo randomly twitching at times)
Pictures of our MIT App inventor app, controlling our robot.
STEP 4: Completed project
In the begining, the group was quite sure that building a project, as simple as a Bluetooth controlled car, would be easy. This however was proven wrong when the group spent about 4 weeks troubleshooting components that where broken or incompatable with Blynk (Our first option for a controller). It also did not help that blynk somehow shortcircuted/broke our original H-bridge. A word of advise from our group would be to avoid Blynk, and use MIT app Inventor instead (Might look complicated at first, but in the end easier to manage than blynks coding requirements).
Our original intent was to first build a car that could be remote controlled and put on Autopilot and then add on features, since in the begining it did not seem to be that hard to implement. But as we countered problems with those functions, we never where able to implement functions such as: Always on "radar scanning" that prints out into console screen on app (Example: https://www.youtube.com/watch?v=kQRYIH2HwfY).
Also the gruop would like to note that using Ultrasound sensors in a moving project caused alot of headache, since sensor readings could fluctuate alot during movement (+ Diffrent surface materials would dampen the ultrasound waves).