A multicamera tool that ensures safer
operations using machine learning
A multicamera tool that ensures safer operations using machine learning
ASIMOV is a self-contained detection system that identifies and tracks humans in the vicinity of ground vehicles. ASIMOV is designed to differentiate between friendly personnel, bystanders, and hostile actors. These inferences are made from a combination of visual cues and human behavioral detections, including adversarial actions meant to mislead common detectors.
Containing a combination of high-performance electro-optical (EO) and long-wave infrared (LWIR) sensors, ASIMOV can be easily mounted to most vehicles and provides 360-degree observations around the vehicle. ASIMOV is self-contained, requiring only an additional power source. All processing is internal to ASIMOV, and the system can be easily transferred between machines.
“We are developing fast, robust technologies for reliably detecting and tracking people in challenging environments and situations.”
Dr. Madison Clark-Turner,
Robotics Scientist and Principal Investigator on ASIMOV
Robust detections even in challenging conditions
Vision systems are challenged by rural or combat environments. The ability to perceive may be negatively impacted by poor lighting, inclement weather, and the presence of visual occlusions such as dust and vegetation. ASIMOV’s use of non-emitting multispectral sensors overcomes all these conditions, allowing for robust detections in circumstances that would otherwise prohibit function.
ASIMOV is explicitly trained to detect and track humans in atypical poses (crouching, sitting, prone) and who are heavily obscured (by foliage or debris). ASIMOV’s detection technology is built on advances from Awarion™, Charles River’s Autonomous Lookout System, which has been successfully integrated and demonstrated in Navy and Army platforms.
ASIMOV’s detections are provided to the user or downstream architecture with metrics that indicate the likelihood that the detection is a false positive/false negative. Relay of these metrics can inform changes in behavior according to the mission, ensuring more conservative operations where human presence is less certain.
Observations made by ASIMOV use non-emitting sensors, ensuring that vehicles are not compromised in their ability to maintain a low profile during military operations.
Relative geospatial tracking
ASIMOV tracks the relative geospatial location of humans as they move about the vehicle, providing actionable information that can inform navigation behaviors that ensure safe vehicle operations and data analytics for reconnaissance missions.
Relative geospatial tracks are informed by vehicle movement, enabling more accurate track maintenance and recovery of lost tracks should the line of sight, to a target that is being tracked, be broken.
Get the ASIMOV brochure
This material is based upon work supported by the ASA(ALT) SBIR CCoE under Contract No. W51701-22-C-0062. Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the ASA(ALT) SBIR CcoE.