PAVBot (Pathfinding Autonomous Vehicle)
Faculty Mentor Name
Joel Schipper, Luis Felipe Zapata Rivera
Format Preference
Poster
Abstract
As machine learning models (MLM) and large language models (LLM) become more prevalent in day-to-day lives, with examples such as ChatGPT and Microsoft Copilot, so does their contribution to autonomous vehicle exploration. In the last couple years, Waymos have populated cities such as Phoenix and provide a new mode of transportation for any person. PAVBot enables further research contributions to the field of robotics and machine learning, which is at the forefront of society’s consciousness in a multitude of ways. The team’s vehicle will navigate using computer vision through LiDAR and two cameras, with data being run through a Jetson Nano and supporting navigation devices. In addition to developing the navigation system, the chassis and power distribution will be designed and built by the team.
MLMs have facilitated tremendous breakthroughs for Autonomous Driving Systems and supports autonomous driving as more research and techniques are developed. This field has three major areas: application, background, and model-oriented surveys. Application-oriented reviews introduce LLMs and MLMs in areas such as prediction, planning, and control while analyzing prompts and learning through reinforcement. Application-oriented reviews are heavily tied to object recognition and proper detection to make correct decisions based off the data analyzed. Furthermore, using LLMs and MLMs in this field allows for the autonomous vehicle industry to become more safe, reliable, and efficient. However, lack of longevity in this field is holding it back, and further research and development will enable these technologies to revolutionize the future of driving. Major competitors have staked a claim in this industry such as Vehicle-to-Everything (V2X) communication, Tesla releasing their Full Self-Driving (FSD) Beta program in 2020 and allowing drivers to test Automated Driving Systems capabilities, Waymo launching their taxi service in 2018 providing than a hundred thousand rides per week, and even Baidu’s autonomous driving travel service that started in 2020 has provided almost a million rides. Further testing, research, and integration in this field is what is holding it back due to the nature of such a concept. Contributions to this area, no matter how small the degree or conditions, are welcomed as test cases. Thus, the creation of PAVBot to navigate a unique driving scenario for validation is a worthy endeavor to aid in the understanding how LLMs and MLMs work and function.
PAVBot (Pathfinding Autonomous Vehicle)
As machine learning models (MLM) and large language models (LLM) become more prevalent in day-to-day lives, with examples such as ChatGPT and Microsoft Copilot, so does their contribution to autonomous vehicle exploration. In the last couple years, Waymos have populated cities such as Phoenix and provide a new mode of transportation for any person. PAVBot enables further research contributions to the field of robotics and machine learning, which is at the forefront of society’s consciousness in a multitude of ways. The team’s vehicle will navigate using computer vision through LiDAR and two cameras, with data being run through a Jetson Nano and supporting navigation devices. In addition to developing the navigation system, the chassis and power distribution will be designed and built by the team.
MLMs have facilitated tremendous breakthroughs for Autonomous Driving Systems and supports autonomous driving as more research and techniques are developed. This field has three major areas: application, background, and model-oriented surveys. Application-oriented reviews introduce LLMs and MLMs in areas such as prediction, planning, and control while analyzing prompts and learning through reinforcement. Application-oriented reviews are heavily tied to object recognition and proper detection to make correct decisions based off the data analyzed. Furthermore, using LLMs and MLMs in this field allows for the autonomous vehicle industry to become more safe, reliable, and efficient. However, lack of longevity in this field is holding it back, and further research and development will enable these technologies to revolutionize the future of driving. Major competitors have staked a claim in this industry such as Vehicle-to-Everything (V2X) communication, Tesla releasing their Full Self-Driving (FSD) Beta program in 2020 and allowing drivers to test Automated Driving Systems capabilities, Waymo launching their taxi service in 2018 providing than a hundred thousand rides per week, and even Baidu’s autonomous driving travel service that started in 2020 has provided almost a million rides. Further testing, research, and integration in this field is what is holding it back due to the nature of such a concept. Contributions to this area, no matter how small the degree or conditions, are welcomed as test cases. Thus, the creation of PAVBot to navigate a unique driving scenario for validation is a worthy endeavor to aid in the understanding how LLMs and MLMs work and function.