UIUC ME 461 Frequency Controlled Segbot

Are you tired of controlling a robot with a controller? Do you wish you could use sounds to control it instead? Well now you can! Mr. Segbot, a TI C2000 MCU controlled two-wheeled self-balancing segbot, uses a Adafruit AGC Microphone Amplifier to take in frequencies he hears to make left/right turns and to move forwards.

Segbot in Action:

Howitworks:

Mr. Segbot is a two-wheeled self-balancing robot created in UIUC's ME 461 course. His brain is powered by Texas Instrument's C2000 F28379D Launchpad and a custom green-board designed for the course. Using a MPU-9250 inertial measurement unit(IMU), Mr. Segbot is able to take the gyro and acceleration values of the position he is at from the IMU in order to balance himself on his two wheels. He does this by using the SPI to send the information to the Dan-777 chip which then transmits the data back in a readable format. Then, using the data received back and a PID algorithm, he is able to stay up-right even someone bumped into him. However, unlike his counterparts in the ME 461 course, Mr. Segbot has the added functionality of moving based on different frequencies he hears.

How does Mr. Segbot differentiate the different noises in his environment? First, he uses his Adafruit AGC Microphone Amplifier to listen to his environment. He then utilizes three 70th order band-pass filters to only pass frequencies within a certain range. He has specifically three filters because he only understands three commands. Turn left, turn right, and move forward. The band-pass filters allows him to filter out the surrounding background noise and thus allows other people to control him by sending frequencies within the desired frequency ranges. At 1100hz, the Mr. Segbot will slowly turn left. At around 750hz he will slowly turn right and finally at around 300hz, he will slowly move forwards. When the desired frequency is heard, Mr. Segbot sends commands through his system to his PID controller such that he can perform the desired operations. Due to Mr. Segbot traveling 2000+ miles to go home for the holidays, he injured his microphone amplifier and as a result couldn't hear well even with his band-pass filtering. This caused some jerk in his movement as can be seen in the video above.

Sensorsused:

Credits:

Thanks to UIUC's Dan Block for creating the ME 461 curriculum and designing the green board circuit used for creating the segbot.

Thanks to UIUC's TAs Chad and Hang for helping with debugging the segbot through out this semester.

Thanks to Mathwork's Yorihisa Yamamoto for modeling and creating the equations of motions of a LEGO segbot which our segbot is based off of.