Power Distribution

We designed a voltage regulator circuit to provide our various circuits, sensors, and actuators with the appropriate voltages. We used 9V batteries for circuits with low current demand. For high current circuits like our H-bridges and servo motors we used 6 AA batteries in series to create a 9V power source. This circuit was built on the same board as our microcontroller.

This is an approximate power and ground diagram for all of our robot's systems.

Close

Microcontroller

All of our software is written in C++ on an STM32F103C8 Blue Pill board which has a 32-bit CPU and an ARM Cortex M3 Microcontroller making it superior to Arduino boards. The blue pill is able to drive motors using PWM signals, receive analog inputs from our reflectance sensors using ADC pins, and receive digital inputs from our other sensors (sonars, encoders, digital reflectance sensor).

Pinout diagram made using KiCAD

We wired and soldered a PCB with our microcontroller, regulator circuit, encoder circuits, and reflectance sensor circuits. We placed the circuit in a aluminum lined box for shielding from electromagnetic noise and wired up a switch which turns on power to the microcontroller and the motor circuits simulataneously.

Close

Drivetrain

Dual H-Bridge Circuits, optically isolated from the microcontroller circuit

This is the H-bridge circuit we designed for driving both wheels in either direction. The design features optocouplers so that the noise from the motors will not effect the sensitive microcontroller. PWM outputs from the MCU are optocoupled and then amplified through a gate driver to drive the motors.

This schematic shows the setup for one motor. The full circuit we built has a second motor which shares the same gate driver.

This circuit was soldered onto a board and secured onto the bottom of the chassis.

Close

PID Tape Following

Reflectance Sensor Array.

To follow black tape, we used an array of 5 reflectance sensors. Each reflectance sensor consists of a IR LED and a phototransistor which we can use to input an analog value into the MCU. Since there is a drastic difference in reflectance from black tape to a light-colored floor, we can determine whether or not a sensor is on black tape.

Using an array of sensors allows us to generate more position states of the robot relative to the tape. This helps us to tune our PID more easily and makes our robot much less likely to drift off of tape.

This is the circuit we use to feed analog values into the Blue Pill:

To be sure we could make all the required turns we connected a digital tape sensor on the side of our robot which would read the small tabs of black tape at the corners of our tape path. When this sensor reads the black tape, our robot will execute a turn.

At this point we were out of ADC pins on the Blue Pill so we made a comparator circuit to turn the analog values into a digital value by comparing it to a value set by a potentiometer.

Close

Precision Driving

Photointerrupters and Motor Encoders

Early in our design cycle, we realized it was very difficult to make our robot drive perfectly straight (without tape) or execute accurate pivots. The motors are controlled by analog values and their speeds are often affected by battery strength. To move precisely, we installed photointerrupters and an encoder on the motor. By counting the pulses from the photointerrupters we are able to determine how much each wheel has turned. This allows us to pivot accurately, drive forward off of tape, and back up the same amount.

Close

IR Beacon Detection

Amplification and AC Coupling Circuit

Although we scrapped our IR detection strategy for our final design, we had built and tested this feature and felt it was worth highlighting.

The idea behind this feature is to be able to navigate towards an IR LED that is oscillating at 1kHz. We chose to detect an oscillating signal so that we can distinguish the presence of IR signal from ambient light in the room

To find the beacon we built a circuit with a phototransistor that receives the signal, amplifies it, AC couples, and steps down the voltage to below 3.3V to be safely input to the Blue Pill.

This is the circuit we designed and built for this application, drawn in Altium.

Once the analog signal enters the Blue Pill, we wrote software to process the signal. We used a cross-correlation algorithm which outputs a value indicating how closely the analog signal matches a 1kHz reference wave. Code for this specific application can be viewed here

This is the circuit we designed to create an 1kHz IR Beacon.

Close