Robotic Systems
Argo Security Bot
As the first class in the robotics series, the requirements for this project were slightly less intense than the future two classes. This project simply involved developing a robotic system or adding simple modifications to a previously built robot to achieve a goal. For this I chose to implement a voice activated home security robot capable of intruder detection, fire detection, and flood damage detection. This project represented much of my first introductions into hands-on robotics and sensors, other than labs, as well as laser cutting, 3D printing, and circuit design.
The robot, nicknamed Argo after Argus the hundred eyed giant of Greek mythology, was an ambitious project for a first-time robotics student. Good chunks of its function were successful though, despite parts difficulties and time constraints. Argo was the first introduction to 3D modelling, design and printing for me on a larger scale. The large blue piece used to house the camera, PIR motion sensors, and indicator LED was custom designed using SolidWorks to house the sensors without the need to truly attach the parts. The chassis was also 3D designed to fit the robot and then cut out of acrylic using a laser cutter
As mentioned above, the robot was to be voice activated for five separate functions that induced separate actions tracked to individual commands. These commands, in the image to the right, allowed the robot to illuminate a hallway with high powered LEDs, server as a party light show, sit idle, line follow a path while searching for intruders, or randomly roam while searching for intruders.
With the need to use cheap parts as a college student, issues were bound to arise. This happened with the voice detection module which unfortunately had to be scrapped and replaced with a more reliable IR remote based mode switching. Also broken upon arrival were the SD card reader/writers preventing the use of the camera and image storing during motion detection. During testing however, one image taken broke through as a corrupted image hinting that the camera was not the issue. However, in light of these issues, improvements were made to the system. Lidar sensors were added for range finding in addition to the forward-facing ultrasonic sensor. The microcontroller was separated in order to improve processing speed and create easier parallel processing with the images, the line sensing, the motion sensing, and power management. Also added was a power cutoff circuit in order to turn of Argo from a distance or if power draw was too high from the battery for expected ranges.
Argo had several of his functionalities at the end but time limitations, scope reductions, and parts malfunctions led to several functions being removed. What remained however, functioned rather well. Using a PID controller on the five-point IR line follower sensor allowed extremely fine control of line centering during the sentry mode implementation. This, paired with a calibration sequence at startup, added flexibility to the line color useable by Argo as the PID controller used contrast, not exclusively black vs white, to determine the path. This did rely on Argo being centered on the line during startup. The lighting functions worked even better than expected thanks to the individually addressable LEDs in the LED strip. The IR remote mode switching, and the PIR sensors also functioned as hoped with LEDs triggering on the side where motion was detected successfully 100% of the time.