ViDAR revealed, Shield AI’s new wide area surveillance pod

Shield AI revealed limited details of its new ViDAR (Visual Detection and Ranging) airborne surveillance pod designed to enable passive reconnaissance and surveillance in a 3rd of April press release. 

ViDAR is designed to be carried by a Group 3 unmanned aircraft system (UAS) – this basically means UAS weighing more than 600 kg with a medium-altitude flight profile, the group stops short of larger systems like the MQ-9 and could be considered a brigade-level asset or similar. ViDAR can also be carried by small fixed and rotary wing crewed aircraft, the company adds. 

The pod is a 60 cm tube designed to be carried on the underbelly of a craft, the image released by Shield AI shows four lens apertures at the front of the pod as well as two on the side angled downwards, that are presumably mirrored on the other side for a total of eight. The camera payload is described as multi-spectral, which indicates a system that can sense different elements of the electromagnetic spectrum (EMS), from optical imagery to thermal imagery. The outputs from cameras operating in different sections of the EMS can be stitched together to provide a more accurate understanding of the terrain and targets. 

ViDAR also carries an inertial positioning sensor as well as a computer running Shield AI’s artificial intelligence. The company does offer Tracker, a moving and stationary target tracking system that uses AI to detect and track targets within full motion video. It does so without requiring the operator to move a gimballed camera or zoom in, enabling multiple targets to be tracked and monitored within the same frame. It can detect objects that are as small as 2×2 pixels, and is five times more sensitive than a human observer, according to Shield AI. 

The AI carried by ViDAR is not specified, however, it is likely that at the very least it draws from the company’s Tracker. ViDAR, “Detects, locates, and classifies a wide range of hard-to-spot, high-value land and maritime threats, including moving dismounts, vehicles, stationary targets, dark vessels (boats with AIS turned off), fast boats, illegal fishing boats, patrol boats, semi-submersibles, and larger vessels,” Shield AI states. 

The passive element of detection is emphasised in the press release with Christian Gutierrez, Vice President of Hivemind Solutions at Shield AI stating, “as modern battlefields become more contested, platforms must detect, locate, classify, and track threats without relying on active emissions.” While this is certainly true, it is less clear how a Group 3 UAS fits within this paradigm as they would likely be detected using other systems because of their proximity to the frontline. However, the press release proceeds to state that ViDAR is designed to complement radar, detecting targets with a low radar cross section. 

Calibre comment

It has proven difficult to find visual detection ranges for systems fitted to Group 3 UAS, however, one specifications page for a system that is likely less capable than ViDAR indicates a maximum visual detection range of 45 km. For comparison, the I-MASTERTM GMTI/SAR Radar from Thales, a smaller system, provides a detection range in excess of 27 km and can be used to survey an 800 square kilometre area in an hour. Other systems on smaller UAS like the Jump-20X from Aerovironment report a detection range of 20 km.

However, radar-based reconnaissance varies based upon the power available to the radar and its aperture – systems weighing 50 kg, compared with the 30 kg Thales system, can detect targets out to 370 km. It is also helpful to think of the relative risk of a radar-emitting platform. The range of smaller radar reconnaissance systems will mean that UAS using them are likely within detection range of a capable adversary. In this case, the drone’s size and radar cross section might work against it as much as the emissions from its radar. The same would likely be true of a UAS using optical detection. Moreover, radar detection is an all-weather capability, it is used to penetrate cloud cover and often complements EO/IR reconnaissance, whereas a camera-only pod would have to descend below cloud cover or call a mission off in heavy cloud cover. 

With all of that said, the ViDAR does seem to offer many valuable capabilities with its AI-enabled wide area surveillance. This enables operators to maintain good situational awareness during complex scenarios, which has often been a challenge with conventional sensors. Drone operators have the option of either seeing in detail, ‘through a soda straw,’ which means they might miss other things that are happening outside of the focus of their sensor. Or, they can see a broader picture without the detail. Combining multi-spectral sensing with AI should enable a better understanding of the terrain and environment as well as allowing focus to find details and conduct more in-depth understanding of identified targets. Similar approaches are being pursued across the defence industry, the ASELFLIR-500 from Aselsan, for example, also comes with built-in AI for target recognition and detection. The general drive to improve reconnaissance and detection from UAS is clear and likely aimed at enabling operators to maintain awareness during long flights with quickly changing dynamics.