Modular Hardware and Software System for Multi-Sensor Environment Perception (MOSEP)

Abstract.

Dynamic environmental processes, from landslides and avalanches to river channel migration and glacier calving, demand innovative monitoring solutions that offer both high spatial and temporal resolution while being robust and cost-effective. Recent advances in robotics and autonomous systems have introduced a new generation of perception sensors, including lidar, radar, and cameras, that have the potential to transform environmental monitoring. While these technologies have long been applied in science, they have traditionally been highly specialized, costly, and designed for narrow use cases. This paper investigates the capabilities of emerging low-cost, lightweight perception sensors for environmental  monitoring, including meteorological applications. We introduce the Modular Hardware and Software System for Multi-Sensor Environment Perception (MOSEP), an adaptable and portable hardware platform complemented by a fully open-source software stack that enables time-synchronous collection of multi-sensor data. MOSEP integrates automotive-grade perception sensors with meteorological instruments, offering a temporal resolution of up to 20 Hz and a range of several hundred meters. The system's robustness and flexibility were validated in real-world scenarios, including stationary deployments and mobile data acquisition on land and sea, such as for iceberg mapping in East Greenland. This paper details the design, integration, and capabilities of the MOSEP platform, showcasing its scalability and adaptability. The results underscore the utility of MOSEP as a versatile tool for environmental sensing, with broad applications in geoscience and other fields. By providing a reproducible, open-source platform, this work aims to motivate researchers to adopt these emerging sensor technologies for their own applications.