Autonomous EV Go Kart
Braking and Steering Systems
Autonomous EV Go Kart
Braking and Steering Systems
Triton AI is a student organization based in the University of California, San Diego, and focuses on learning AI and competing in robotics competitions. Among its initiatives is an autonomous electric go kart, which competes against 9 other teams nationwide in the Autonomous Karting Series.
In 2022, the Triton AI Autonomous EV Go Kart placed 2nd, with 1 out of the 5 laps completed, and the fastest lap time as 2 minutes and 44 seconds!
The focus on this project was on the development of the hydraulic brake system with the goal of reducing rear wheel skidding, and the installation of a speed encoder onto the rear axle of the go kart in order to track and monitor the action of the wheels skidding.
Another task that was originally assigned was the integration of a system in which human induced torque on the steering wheel resulted in the override of the autonomous driving mode. While this portion was researched, the sponsor and team decided to not move forward with this task in order to focus on the wheel skidding matter.
In regenerative braking, the motor causes the direction of rotation of the axle to reverse. This causes a change in direction of the wheel rotation - if the wheels were initially moving forward, they are now moving backward - which led to the go kart's loss of control. The go kart loses control similar to an oversteer situation, which is characterized by the rear end sliding outward in relation to the front wheels during a turn. It drifts, going into a sideways skid causing a loss in velocity, momentum, vehicle control, compromising traction and overall race performance.
While the team had a hydraulic brake system bought, it remained unused due to challenges with installation. In the context of racing, by utilizing a hydraulic braking system instead of regenerative braking, the vehicle stability is maintained as it will no longer lose control on deceleration. Hydraulic brakes apply braking torque to the axle which then slows both wheels down and maintains stability.
Regenerative braking diagram of motor.
UTILIZE a brake system based on fluid pressure control
CHARACTERIZE the braking force required to decelerate the kart smoothly.
IMPLEMENT a method that measures wheel speed on rear axle of go kart
ENSURE precision and accuracy in the speed measurement through the mounting designs
CREATE a system in which autonomous driving can be overridden by human touch
RESEARCH & INSTALL sensors for autonomous steering torque override
It is also important to note that all electronics (such as sensors) must output a signals less than 3.3V in order for it to be compatible with the go kart's RTC (Real Time Controller). Also, all aspects of the design must be easy to replicate, and selection criteria for purchased parts must be based on accessibility in order for other autonomous go kart teams to be able to recreate the design at low costs and purchase these parts from convenient sellers such as Amazon, eBay, and McMaster Carr.
The following final designs for our tasks are summarized below. For a more detailed analysis, please visit the Final Design tab.
Past project team members installed a hydraulic brake system that contains the following components: the brake pedal, a master cylinder piston, calipers, brake pads, and brake fluid that travels through the brake line. The team was able to utilize the hydraulic brake system using a brake caliper and rotor setup, which applied braking torque on the rear axle and allowed the car to slow down or come to a complete stop without skidding. The team was able to quantify the maximum braking torque that must be applied to avoid skidding under heavy braking or the wheels locking.
The final design of the hydraulic brake system incorporates the linear actuator to be mounted higher to gain a mechanical advantage, a clevis arm, and designed master cylinder arm. The following are CAD images of the final designed linear actuator mounting plate and the designed master cylinder arm with the clevis arm attached.
The actuator generates 90 lbs of force, and remounting it 2 inches higher had created a larger moment arm in order to increase the braking force applied.
A clevis arm, threaded to fit the end of the linear actuator, was added, and ensured reliable connection and consistent contact as the brake pedal angle changes with the actuator’s movement.
A new brake master cylinder was designed to provide adequate and constant surface area contact throughout the actuator’s range of motion.
Diagram of the brake set up before and after redesign.
The speed encoder system consists of an ABS reluctor ring and a hall effect sensor. The hall effect sensor works by disrupting the magnetic field from the ABS ring, resulting in the output of a digital signal that can be processed and converted into speed data.
The parts selected were an ABS ring from Jeep Wrangler 2007-12 models and a Littelfuse Inc. Digital Hall Effect Sensor. The inner diameter of the ABS ring was significantly larger than the diameter of the axle, so a mount was created in order to allow it to fit onto the axle without spinning independently. A mount was also designed and created to be installed onto the go kart chassis, with 2DOF adjustability to allow for precise readings of the ABS ring.
CAD of the sensor mount (left) and ring mount (right)
Images of installed sensor (left) and installed ring (right)
Final installation onto the go kart axle of ABS ring and press-fit mount, hall effect sensor, and sensor mount.
Prior to the redesign of the brake system, the maximum brake pressure obtained was 137 psi and brake pressure was only displayed in the last 12.5% of actuator displacement as shown in the left. The relationship between the linear actuator and brake pressure was not gradual as required -- this meant that you would have to apply more force to the brake pedal in order to achieve the desired braking effect.
After the redesign and installation of the brake system, there was a gradual relationship between actuator displacement and brake pressure, with a maximum pressure of 927 psi, which is approximately seven times the maximum pressure obtained with the initial Top Kart purchased system. The results demonstrated that the redesigned system produced a mechanical advantage with consistent contact between the master cylinder arm and actuator, leading to improved brake performance.
Relationship between brake pressure and linear actuator displacement, before (left) and after right) redesign.
Skidding occurs due to regenerative braking and car reverses back
No skidding and car comes to a complete stop once brakes are applied
After finalizing and machining both the three set screw shaft collar and the 2-DOF sensor mount, they were securely mounted onto the go kart, ensuring the correct distance between the sensor and ABS ring. Testing was then conducted to confirm that all teeth were being accurately read by the sensor. With the finalized mounts and correct distance between the ring and sensor, all teeth were successfully detected at high and low speeds without the presence of missing waves in the signal.