Project number
27025
Organization
Sandia National Laboratories
Offering
ENGR498-F2026-S2027
Future autonomous systems may need to operate in contested environments where GPS is unavailable and radio-frequency emissions are restricted. This project will develop and demonstrate a small uncrewed aircraft system that can perform a search-and-mark mission using only onboard sensing, computation, and autonomy. The UAS will launch from a known location, navigate without GPS, search for an object of interest, identify the object, illuminate it with a pointable eye-safe collimated light source, and return home without intentional RF emissions during the mission. The object of interest will be a life vest placed at least 200 m from the launch site within a search region of at least 30" m"×30" m" , and it will not be visible from the launch point. During the mission, operators may observe the system but may not provide input, commands, navigation updates, or object-location cues to the UAS. Mission execution should be performed autonomously after launch.
Student Teams Will
• Design or integrate a UAS platform capable of autonomous operation without GPS-based navigation.
• Develop onboard sensing with onboard or off board estimation methods for navigating relative to the launch point.
• Implement autonomous search behaviors for locating a life vest in an area not visible from launch.
• Develop computer vision or other sensing algorithms to detect the life vest, compute can either be onboard or communicated through a non-RF link
• Integrate a pointable eye-safe collimated light source to illuminate the life vest after detection that can dwell on the life vest for an amount of time prescribed at the start of the mission ranging from 10 seconds to 5 minutes.
• Estimate the life vest location relative to the launch point with less than 8 m error.
• Demonstrate autonomous mission execution without operator input during flight and without intentional RF emissions.
Student Teams Are Encouraged to Pursue
• Providing mission status or operator feedback while maintaining the no-RF-emission and no-operator-input constraints.
• Implementing SLAM, visual-inertial odometry, optical flow, or other GPS-denied navigation methods.
• Improving autonomous search efficiency and return-to-home accuracy.
• Operating in cluttered environments or against cluttered visual backgrounds.
• Operating in degraded visual environments, such as low light, glare, dust, haze, shadows, or partial occlusion.
• Improving object detection confidence and reducing false detections.
• Demonstrating robust performance over multiple trials with varied life vest placement.
Student Teams Will
• Design or integrate a UAS platform capable of autonomous operation without GPS-based navigation.
• Develop onboard sensing with onboard or off board estimation methods for navigating relative to the launch point.
• Implement autonomous search behaviors for locating a life vest in an area not visible from launch.
• Develop computer vision or other sensing algorithms to detect the life vest, compute can either be onboard or communicated through a non-RF link
• Integrate a pointable eye-safe collimated light source to illuminate the life vest after detection that can dwell on the life vest for an amount of time prescribed at the start of the mission ranging from 10 seconds to 5 minutes.
• Estimate the life vest location relative to the launch point with less than 8 m error.
• Demonstrate autonomous mission execution without operator input during flight and without intentional RF emissions.
Student Teams Are Encouraged to Pursue
• Providing mission status or operator feedback while maintaining the no-RF-emission and no-operator-input constraints.
• Implementing SLAM, visual-inertial odometry, optical flow, or other GPS-denied navigation methods.
• Improving autonomous search efficiency and return-to-home accuracy.
• Operating in cluttered environments or against cluttered visual backgrounds.
• Operating in degraded visual environments, such as low light, glare, dust, haze, shadows, or partial occlusion.
• Improving object detection confidence and reducing false detections.
• Demonstrating robust performance over multiple trials with varied life vest placement.