University of Arkansas College of Engineering Arkansas Space Grant Consortium Autonomous Vehicle Challenge Proposal Period of Performance: Fall 2023 – Spring 2024 Submitted by Primary Investigator: Phone: Robert Saunders 479-575-6047 Email: rsaunder@uark.edu ASGC Representative: Dr. Larry Roe Research Sponsor: Dr. Eksioglu Project Description University of Arkansas EE Department 1 The Arkansas Space Grant Consortium Autonomous Vehicle Challenge has undergone a significant change in rules to alter the method of completing the obstacle course. By introducing an unknown environment that alters the positions of obstacles teams must create new solutions to complete the course in the fastest time possible. Previously an artificial intelligence dominated the competition by recreating the movements of an operator that had run the course in the practice rounds. Due to the introduction of moving obstacles has made this solution obsolete. The combined efforts of the Electrical Engineering, Computer Science and Mechanical Engineering teams have come together to propose a new solution to handle the new challenge presented. The following description will include the project goal along with a brief description of each objective, then the relevance of the project and how it relates to the NASA Mission Directorate. Taking into account an unknown changing environment our team has decided to use mmWave radar and a small modular camera to navigate the obstacle course. Our team’s main goal is to identify the course obstacles prior to the race, and creating a planned route to complete the course in the shortest amount of time. To do this several objectives have to be completed such as: constructing a sturdy vehicle, use a mmWave antenna and camera to send information to a central computer, parse through the data to determine the location of objects, decide how to act on the information, and then translate the instructions to the motors and steering to complete the course. Breaking the objectives down further the Mechanical Engineering team will consist of four undergraduates who will design a frame, suspension, and steering. The team that competed last year experienced a large difficulty completing the ramp obstacle due to accrual of damage over time. To combat this a large emphasis is being placed on handling the fall and speed at which the ramp is jumped. The materials will be partially salvaged, and 3D printed to create a light but sturdy vessel for the electrical and computer components. The body design is currently a work in progress in order to maximize the efficiency of the radar and camera to ensure the most accurate data collection. To steer the vehicle a two-system motor setup is devised to allow for greater speed and control using pulse width modulation (PWM) to control the angle of movement. The motors currently devised will be a set DC/DC brushless motors with a speed controller built in for ease of use. University of Arkansas EE Department 2 With a car design in progress the ability to find the objects is the next objective to conquer. With the obstacles being given a +/- 15ft change the idea to use radar as our primary source of input data was devised. To overcome this challenge three Electrical Engineering undergraduates will be tasked with using radar. By using the distance sensors, the Razr-Car could be placed into any environment and be able to locate and maneuver around the obstacles on the track. As opposed to the idea of AI integrated runs our vehicle should be able to complete any variation of the sample course provided. After the vehicle knows where the object is a camera can be used to identify the color of the obstacle to help decide the direction and route the car needs to travel. The radar that will be used is a mmWave development board (ICWAVEBOOST) and antenna package (IWR6843AOPEVM). The data from the radar will be sent to the Computer Science team’s computer via a USB cable utilizing UART communication so that the data can be parsed through. A modular small camera with depth perception and RGB capabilities will be plugged directly into the nano computer. To power the devices a battery will be used with converters to properly power each device to its set specifications on the controller board. Now that there is a car and input data for the location of objects the next objective is to decode the data into usable information and decide what action for the vehicle to take. A team of five Computer Science undergraduates will be tasked with using the information provided and deciding what action to take. To do this a small computer in our case, a Jetson nano development board will be used with an object-oriented AI identification protocol that will differentiate the obstacles. The radar will give a distance to the obstacle and once within a set distance the computer will use image recognition to decide which action will be taken. A preliminary idea is that every blue obstacle requires a right turn, the yellow obstacle a left turn, red obstacle being through the middle and ramp being straight over. After the computer has decided which action to take will transmit simple commands back to the controller board used to manipulate power to the devices and control the motors. The simple actions currently consist of percent left, right and forward. Now that there is a car, input data and instructions from the computer the vehicle can begin to move. Since the two motors are controlled by PWM the signal transmitted to each motor will result in a variable direction to travel. Since the radar and camera will be constantly sending data the vehicle will be able to move appropriately around the obstacle and be able to find and move University of Arkansas EE Department 3 towards the next obstacle. The speed of the vehicle will be determined as well by the computer using the distance from the next obstacle and which obstacle was completed most recently. To test the car a small-scale course will be constructed at the nearby car garage to see if the vehicle can navigate around the objects as intended. Doing so will allow for variable conditions to be tested for tuning of the camera or motor speed. Since this is competition with previous knowledge of competitors, cameras are liable to be inaccurate ate due to lighting conditions and the computers are limited by action speed. This enables the vehicle to be fine-tuned and ready for competition post course simulations and diagnostics. The question of relevancy on behalf of NASA or the university was a large topic when it came to deciding how to locate and avoid obstacles. Our team settled on radar and cameras after discussing how a rover could locate objects on the moon or mars. By utilizing radar, the rover would be able to know the location of various objects around them and navigate towards the objects. Once close to the object a camera or other sensors could be used to then decide how to act upon the obstacle be it navigating around or taking a sample. The vehicle made in this autonomous vehicle challenge will be made in the same light as being placed in an unknown environment and being cable of selfnavigation and decision-making. For the relevancy of our staff members, Dr. Eksioglu specializes in transportation and logistics. When developing an algorithm to detect objects and choose what to do in various situations will be of invaluable help. With the help of Dr. Larry Roe our ASGC representative the knowledge of space and planetary sciences can be taken into consideration for a vehicle that may be place in hostile environments and succeed. By combining the knowledge of transportation algorithms and science of space the implementation of radar and camera recognition will further the research of autonomous maneuvering and unpiloted vehicles under the Exploration Systems directorate specifically Exploration Ground Systems. The primary investigator and lead figure in helping decide on radar technology is Robert Saunders who will help guide the teams along with Jason Bailey for the Mechanical Engineering team and Matthew Patitz for the Computer Science team. The professors assisting on the challenge will oversee the designs and work done by the students to make suggestions and verify work done. University of Arkansas EE Department 4 Scope of Work Mechanical Team: Frame 4+ hours weekly Steering Chassis Suspension Spec Motors (with EE help) Electrical Team: Controller Board Schematic 4+ hours weekly Send Input Data to Computer Receive Actions from Computer PCB Layout for Controller Board PCB Fabrication Test and Optimization Computer Science Team: Code Framework 1+ hours weekly Receive Data from Radar/Camera (will increase next Object Recognition Semester) Vehicle Actions Send Actions to Microcontroller Budget Justification As we are applying for a grant from the Arkansas Space Grant Consortium a detailed list of where the money will be spent will be detailed below. A few of the components detailed above in the goals and objectives section are already in possession of the teams. The mmWave radar, Antenna, and Jetson development board are recycled from various other projects done by the Electrical Engineering and Computer Science departments. Though the materials needed to construct the car for the Mechanical team will need to be purchased as well as the two motors that will used. The motors currently being looked into are the DC/DC brushless motors as mentioned above (3982University of Arkansas EE Department 5 DCM-1036-ND) but have not been decided on yet. Then the controller board, electrical components, and PCB will be purchased to enable power and communication among the devices in question by the Electrical Engineering team. The computer science team has not yet picked out a camera that will work in tandem with the Jetson development board so that will also need to be purchased. The majority of the budget will go towards the travel expenses for the various sub-teams (ME, EE & CS) that consist of our University of Arkansas team. With all the undergraduate sub teams combined there are twelve students and three staff members who will need to be driven to the competition event. The travel expenses will cover the renting of buses/vans, food, and shelter for the team. The PI and undergraduate students on the team will not be compensated/salaried using the Space Grant. The team currently has no awarded support an is using the mentioned recycled equipment currently to complete the competition objectives. University of Arkansas EE Department 6