Tracker

Sentient Tracker is a Moving Target Indicator (MTI) enabled by AI that provides moving target indication like a radar target indicator with the additional capability to track targets with breadcrumbs. Sentient Tracker is the AI-enabled version of Sentient’s moving target indicator, Kestrel, which automatically detects moving objects within electro-optical (EO) or infrared (IR) full-motion video from a range of uncrewed and crewed aircraft.

Origins of the Moving Target Indicator

MTI is typically referencing a mode of operation in radars to contrast a moving target such as aircraft from the unmoving clutter. This is done by comparing the doppler frequency shift of the moving object’s received signal from the radar antenna. The Doppler effect is used to measure the phase shift of reflected radio wave pulse in moving objects (the frequency of the reflected pulse changes, while a stationary target with no change is filtered out). MTI stores a single pulse of the received signal for the time between broadcasts (pulse repetition frequency). Modern radar systems including synthetic aperture radar have moving target indication that allow the measurement of radial velocity via the Doppler shift from a single signal.

The probability of detection is dependent on antenna size and the amount of power it radiates. Location accuracy is also dependent on the position of the radar.

Advantages and disadvantages of MTI radar

As MTI radar is one type of pulse radar, its advantages include using very low pulse repetition frequency to avoid range ambiguities and determine target velocity to differentiate moving from still targets. Supposedly easier to use than pulse doppler radar, it is more economical and have a high antenna bandwidth.

Disadvantages of MTI radar are : Blind speed does not get detected by pulse MTI radar. Blind speed is defined as magnitude of target’s radial velocity when moving target appear as a stationary target. A typical MTI radar uses high-pass filter to remove energy at low Doppler frequencies. Since the frequency response of a high-pass filter is periodic, some energy at higher Doppler frequencies is removed and hence these targets will not be detectable by radar. Other disadvantages include doppler ambiguities, requires active transmissions while in use, requires extensive radar training and interpretation, requires cross-cueing to motion video system for target confirmation. The larger the radar the more accurate. The larger radar form factor is not always suitable for a smaller moving platform such as tactical drones. Short antennas affect the target location accuracy when there are multiple target vehicles moving into an area with many roads. There is a minimum detectable velocity (MDV) for Ground Moving Target Indicator. Any target with a velocity less than the MDV cannot be detected because there is no sufficient Doppler shift in its target echo to separate it from the clutter return.

Moving Target Indicator applications

MTI radar can be installed on UAVs (uncrewed aerial vehicles) and crewed aircraft as part of the sensor payloads. The algorithm for signal processing compares radar signal collected from the radar antenna to differentiate the moving targets. The doppler radar also provides precise location of target and tracks movement in real time. As the location of the moving target changes, the distance in relation to the passing mti radar is overlaid on an optical map as detection. This allows the radar system to produce tracks to highlight the path of the target detection. MTI is used in defense and military applications including ISTAR ( intelligence, surveillance, target acquisition and reconnaissance as well as search and rescue. Modern radars have systems that can detect and track ground vehicles with Ground Moving Target Indicator (GMTI). MTI radar is often used to cross-cue electro-optical/infrared (EO/IR) sensors in full motion video gimbal systems. An example of MTI radar is Lynx radar from General Atomics Aeronautical Systems.

Full motion video gimbal systems

Full motion video (FMV) gimbal systems have EO/IR sensors that have a narrower focus, referred to as the “soda-straw effect”. This is the restricted view of land or sea as though one is looking through a soda straw while needing high-resolution imagery. A wider angle is ideal to enhance and speed up target detection for better situational awareness.

What is Sentient Tracker?

Sentient Tracker is an AI-enabled Moving Target Indicator solution. Based on Kestrel, a product with the technology readiness level (TRL) of 9, Sentient Tracker autonomously detects moving targets by augmenting full motion video systems to detect targets in areas that are 20 times larger, eliminating the “soda-straw effect”. The targets can be persistently tracked by adding “breadcrumbs’ to follow moving targets without losing sight of the big picture. The AI analytic for mti processing enhances detection of moving targets and accuracy of the detection. While it is not a pulse doppler radar and does not utilise the transmitted pulse of the doppler spectrum, Sentient Tracker provides real-time situational awareness like moving target indicator radar for a wider range of aircraft that are not always suitable for radar.

2024_04_29_sentient_tracker_before_after_v5.1 (540p) from Sentient Vision on Vimeo.

As radial speed is not a factor, Sentient Tracker detects both fast and slow moving targets including people and vehicles over challenging terrain especially those with tall structures. Sentient Tracker significantly reduces parallax detection from tall structures such as trees, towers and stadium lights, which are common issues with all moving platforms. Sentient Tracker requires FMV gimbals to have minimum processing modules required for industrial-grade edge AI. Enquire with Sentient here.

Capabilities of Sentient Tracker

Enhances mission outcomes for operators tracking multiple targets at once without missing movement in the periphery.
- Increases surveillance areas by 20 times, eliminating the “soda-straw” effect.
- Enhances reliability with increase in number of target detection by up to 4 times.
Improves accuracy of false positives significantly.
- Minimizes parallax caused by moving platform over tall structures to near zero
- Minimizes ‘ghost’ detection of fast moving targets and shadows to near zero
Detects slow or fast moving objects regardless of speed
Improves tracking persistence of targets detected
Passive and does not actively transmission during use.
Accurate yet small form factor addition to full motion video systems, suitable for a wider range of aircraft
More economical than MTI radar

Typical uses for Sentient Tracker

Sentient Tracker combined with full motion video gimbal systems enhances mission outcomes of

Law enforcement
Counter drug trafficking operations
Counter Illegal fishing operations / marine life protection
Counter people smuggling operations
Counter piracy / maritime hijacking operations
Intelligence, Surveillance, Target Acquisition and Reconnaissance
Search and Rescue (ISTAR)

Kestrel moving target indicator

Kestrel is a software that enables detection, tracking and geolocation of moving targets in an area that is 20 times larger than an EO/IR camera’s narrowest field of view, eliminating the “soda-straw effect”. Visual detection cues alerts the operator to investigate slight movement that is invisible to the human eye and targets can be tracked for pattern of life analysis. Kestrel is at TRL9 and operationally proven. Kestrel integrates with many crewed and uncrewed aircraft sensors, mission systems and is complementary to ground control stations and remote video terminals.

Kestrel is the result of more than 20 years of AI / computer vision research and deployment, and detects moving objects, people, vehicles and vessels over challenging terrain in real-time or for forensic operations. Raw data can be quickly transformed into actionable intelligence.

Capabilities of Kestrel

Enhanced surveillance with small object detection down to a few pixels in size
Automatically detecting moving land objects (vehicles and people) or maritime objects (small boats and rubber rafts)
Provides dedicated modes optimized for search and rescue or maritime surveillance.
Extractable data including size, speed and direction of moving target for real time or forensic (pattern of life) analyses.
Moving target tracking with breadcrumb representing visible movement history and predicted trajectory
Geo-location data compliant with STANAG 4609 and MISB 0601 in MGRS and Latitude/Longitude.
Works with ISR systems, existing mission systems, sensors, ground control stations and RVTs
Integration options include Windows application, APIs for Windows, Linux and embedded processors.

Testimonial from a Kestrel user

Kestrel autonomously detects moving targets and highlights them in a motion video, reducing operator fatigue . It enables a wider field of view than the human eye. “Kestrel detects any small movements in the field of view and alerts our operators to objects that are easily missed – camouflaged vehicles in tough terrains, dismounts or insurgents leaving houses. This gives our commanders a complete understanding of the situation on the ground.” WGCDR David Riddel, RAAF Deputy Director – Air Combat Capability