Hostile fire Autonomous Detection and Evasion System (HADES)

Small arms, snipers, mortars, and artillery are primary threats to squad- and platoon-level operations during missions, as well as to small UAS, which are high value ISR assets. UAS are operated beyond line of sight with limited comms, making them vulnerable. Small UAS are often expensive ISR assets that need protection. HADES is a low-SWaP (size, weight, and power) opto-acoustic hostile fire detection (HFD) system for Group 1 or 2 unmanned aerial systems (UAS); it provides a multimodal hostile fire detection sensor that can be tied into the aircraftÆs autonomy software, to support detection of both super- and subsonic hostile fire, giving the vehicle a way to protect itself by taking evasive action as needed. In previous work, SSCI developed a MWIR-acoustic hostile fire detector system for helmet-mounted and Group 1 UAS applications, able to compensate for platform motion and geolocate shooters using sensor fusion across sensor modalities. A system was designed and fabricated using non-imaging single element sensors, with integrated sensors and signal conditioning electronics, processor, inertial measurement unit (IMU), and power management. Initial testing used an array of standard uncooled PbSe sensors in T05 packages, mounted in a flattened, open hemispherical structure. This design will serve as the starting point for the HADES effort, leveraging the peripheral sensors, non-imaging single element approach, and detection, localization, and fusion algorithms. The novel part of the HADES approach uses non-imaging MWIR detectors, placed in a ring that provides wide overlapping fields of view for continual 360░ situational awareness. Non-imaging detectors are cheaper, more robust, passive, respond faster, and can pick up very short duration events, like the flashes from gunfire and mortars. For practical application on small UAS, sensors must have small size, weight and power (SWaP) characteristics. The acoustic sensors for gunfire applications have been thoroughly tested; an array of four commercial piezo MEMS microphones is sufficient to provide good coverage in azimuth and elevation with no ambiguities in detection. Our multimodal approach uses a single low-SWaP core processor board, and fusion algorithms to combine features from the two modalities. The infrared system uses a wide field of view single element uncooled sensor in the mid-wave infrared (MWIR) band, to detect weapon ōflashö. The acoustic system is able to detect the muzzle report (audible sound) of weapon fire, known as the ōbangö, and the bullet wake from supersonic rounds, known as the ōcrackö. The bang and crack are combined to compute an estimate of the bearing to a source of fire, as well as a range to source estimate. These combined sensing approaches offer a unique and robust hostile fire detection solution in orthogonal domains that yields an accurate estimate of the source of fire. This system has a very low false alarm rate, even during challenging conditions onboard a UAS.