LEANSense is a portable, real-time sensor streaming and algorithm evaluation platform developed for the LEAN group’s research in low-energy autonomous vehicles. The system pairs LEANCam, a Swift-based iOS application, with a C++ host library, using a Rerun server as middleware to enable simultaneous live 3D visualization and flexible multi-language API access to streamed sensor data including RGB images, depth maps, LiDAR, and IMU readings. Existing autonomy evaluation approaches such as motion capture and SLAM-based post-processing are accurate but hard to deploy and yield limited data diversity; LEANSense addresses this by enabling rapid, configurable dataset collection across diverse real-world environments with minimal setup. This SuperUROP continues development from Fall 2025 across three milestones: implementing live data query APIs on top of the existing Rerun recording infrastructure, refining system performance through serialization and buffering optimization alongside thorough technical documentation, and finally integrating LEANSense as the sensor input source for Gleanmer.

During this SuperUROP, I hope to deepen my understanding of systems programming and low-level hardware by contributing to a project that sits at the intersection of software development and energy-efficient autonomy. Working on LEANSense will sharpen my C++ skills, strengthen my intuition for system architecture, and give me hands-on experience designing APIs for real-time, resource-constrained environments. Most importantly, I hope to write code that is correct, maintainable, and genuinely useful to the researchers and future students who will build on this work.