The team was challenged to develop a small inexpensive camera system that could monitor a larger area to detect and classify objects in the airspace between 200-300 feet away. The team developed a camera array that provided 360 degree coverage with small wide angle cameras and a single gimbal mounted high zoom camera. The solution allowed the lightweight edge processor to pick out movement on the wide angle cameras and point the high-zoom camera at the object for automatic object detection.
Autonomous Racer Drone
A defense customer was looking at developing a lightweight inexpensive counter UAS solution capable of intercepting a flying object. The scenario required a long range light and inexpensive solution. The team settled on retrofitting racer drones that have excellent performance characteristics with a flight controller that could receive instructions from a central system. The solution used a heavy lift drone to carry and drop the lightweight drones 10-15 miles out, they would then fly and execute their mission while maintaining communication with the larger drone capable of carrying more compute power.
Industrial Sensor Cluster
A manufacturing client required sensorizatoin of their existing manufacturing line made up of legacy conveyor belts and electric motors. Before spending the money to install and connect permanent sensors, they asked us to identify the location and the sensor type that would yield the most value in detecting failures ahead of time. The team developed a battery powered sensors cluster capable of running in harsh conditions for up to 9 months. In addition to the sensors, the team utilized LoRa communication to eliminate the need for extensive dedicated infrastructure.
Heavy Lift Quad-Copter
To support multiple client requests to fly heavy sensors, cameras and edge compute, the team set out to design a heavy-lift quadcopter that could automatically execute flight plans. The requirements included an extended range of up to 15 miles and a carrying capacity of a minimum of 25 lbs. The design allowed for a modular propulsion system that allowed the team to easily convert the quadcopter from a 25 lbs to a 50 lbs payload capacity, all without major changes to the platform.
Emergency Ventilato
During the COVID outbreak in 2020, the team responded to the call for emergency ventilator designs that could be manufactured in large quantities, using readily available parts. The final design utilized a windshield wiper motor, a ventilation mask and a set of industrial sensors to ensure safe operations. To meet the reliability requirements that come with a medical device, the team integrated additional sensors throughout to identify suboptimal operations and alert the operator immediately.
The challenge presented to the team was to develop a network of sensors capable of monitoring a large area for two weeks. The team developed a battery powered (solar cell optional) sensor node made up of environmental sensors capable of picking up nearby activity. These nodes were deployed from a drone over a large area and transmitted their reading over LoRa to a long endurance fixed wing UAV equipped with a LoRa receiver.
The team was tasked with taking a standard golf cart and transforming it into a modular transportation unit capable of moving people and supplies. The upgrade included autonomous navigation capabilities with integrated computer vision sensors. To ensure this unit was able to take whatever the terrain had to offer, the clearance was upgraded and to ensure it could carry whatever was thrown at it, the team fabricated a customer bed and railing system.
The client was looking for a low cost autonomous vehicle to inspect ground based industrial assets as well as monitor for activity, both from unauthorized individuals and animals. The requirements included at least an hour of charge, the ability to execute routine inspections automatically and the ability to detect specific objects within the inspection. The team design and built a small UGV with object avoidance and object detection capabilities capable of automatically executing preplanned missions.