Spherical Sensor Configurations (SSC) have been designed for detecting and tracking signals in three dimensions. The Spherical Sensor Configurations offer distinct advantages over contemporary imaging systems, significantly enhancing three-dimensional (3D) situational awareness. Sensor systems utilizing a spherical geometry as a foundation (Figure 1) can be designed for a variety of applications.
A singular ring of sensors provides a basic device that can monitor sources in a two-dimensional plane. These systems calculate extremely accurate angular directions to signal sources. More sophisticated systems with full spherical sensor placement are being designed for monitoring multiple targets in any spatial orientation. Information generated from these systems can be integrated with contemporary imaging systems for further target identification. Full spherical SSC systems offer a 4 Pi steradian Field of Regard (FOR).
The underlying concept of the innovation is based upon fundamental properties of signal transmission and its reception on a spherical surface. The technology exploits the principle that a source of light will illuminate one hemisphere of a spherical object (similar to the Earth/Sun system). The significance of this concept lies in the consistent and mathematically predictable position of the illuminated hemisphere relative to the source. Harnessing this concept, a spherical receiver can be constructed with strategically placed sensors to determine the position of the illuminated hemisphere and, subsequently, the 3D direction of the source. Spherical Sensor Configurations offer distinct advantages over contemporary imaging systems in monitoring a wide FOR. SSC systems overcome imaging Field of View (FOV) limitations, enabling a single sensor system to view targets in all directions, significantly enhancing 3D situational awareness.
Spherical Detection Systems (SDS) are being designed for detecting and tracking infrared (IR) heat signatures, primarily in the 3 - 5 and 8 - 12-um thermal imaging bands. Current prototype development is in the 8 - 12-um band for detecting and tracking human and vehicle infrared (IR) targets. Analysis is being performed in the 3 - 5-um band for the detection of hotter temperature targets including missiles, RPGs, highspeed vehicles, and arms fire. The SDS can simultaneously track multiple targets in any spatial orientation, making it the ideal sensor system for Infrared Search and Track (IRST). The SDS can be fitted with sensors that can sample into the megahertz and analyzed with an onboard DSP for ultra-high-speed threat detection and tracking. The system offers continuous, passive sensing in a mechanically passive package with low power consumption. The wide FOR and high-speed sensing is well suited for capturing the location of initial launches or other incoming threats.
A Two-Dimensional Prototype
The 2D prototype was designed with 30 IR sensors mounted on a 4"-diameter ring. The sensors have a 100° FOV and are each separated angularly by 12°. The sensors receive light in the 350 - 1150 nanometer spectral range. The sampling rate is variable up to 10 kHz. Sensors directly facing an IR source produce a maximum response from the source relative to the other sensors on the ring. As the angle of the sensor relative to the source increases, the response from the sensor decreases. The sensor results can be plotted in a histogram fashion with the x-axis representing the angle of the sensor on the ring and the y-axis representing sensor response in Volts. The histogram takes on the shape of a Gaussian function.
The center of the peak represents the strongest sensor response and therefore the direction of the source. The sensor data is then fitted with a peak as shown in the display. The maximum of the peak is calculated and the corresponding x-value is obtained to determine the direction of the source. Using this mathematical technique, incoming sensor data is analyzed to determine the direction of the source with better than 0.01° (.174 mrad) of accuracy. Current advancements in circuit design and peak-fitting algorithms should significantly increase the accuracy.
In addition to determining the maximum of the peak, the peak width can be measured to determine the width of a source. The peak width is directly proportional to the angle subtended by a source. The current prototype uses a sensor with a very wide FOV, which allows for about three independent sources to be tracked. By increasing the density of sensors and decreasing the field of view of the sensor, more sources can be tracked simultaneously.
In designing the prototype, a Fresnel lens was used with a 30° FOV allowing for the tracking of more independent sources. Currently, the system uses 2M thermopile sensors to detect human IR in the 8 - 12-um thermal imaging band. Using a combination of a thin Fresnel lens and the 2M detector, field tests have obtained ranges of 100+ feet under average conditions (23°C). The data acquisition and control system is based on a PC- 104 system running XPembedded with two 16-channel data acquisition cards for the analog input. The control program operates at 10 Hz with a similar display to the VIS-NIR prototype. The unit provides a robust method for locating thermal targets for ground- or marinebased applications.
The 2D Human IR System (Figure 2) can be utilized on a ground-based vehicle, tripod, or as a pole-mounted system for detecting and tracking human IR targets. Networked systems can be integrated for wide-area surveillance. Lower-power field-programmable gate array (FPGA) systems are being developed for remote deployment.
Lucid Dimensions has independently developed methods for distributing sensors on a sphere. This method uses a triangular lattice spacing that exhibits spherical symmetry when projected onto an XY plane. Similar to the 2D histogram of the circular prototype, sensor data over the sphere produces a 3D Gaussian histogram. The sensor data has a maximum where sensors are directly facing the source. Sensors surrounding the maximum will fall off in a predictable manner based on the relative angular position.
A wide range of algorithms can be applied to analyze the incoming sensor data. Using peak deconvolution techniques, overlapping peaks can be analyzed for closely spaced sources. These algorithms include baseline subtraction, smoothing, peak searching, and finding peak maximums.
Figure 3 shows the sensor distribution from three different sources. Assuming all sources are equal, the blue source is the furthest from the sphere, while the red is the closest and green is in between. The three images show the effect of decreasing the sensor FOV in order to separate out peaks. This results in less active sensors per target, allowing more sources to be tracked without having to perform extensive peak deconvolution.
The initial 3D spherical prototype is being designed with a 503 modal sensor distribution. The major problem associated with the system is the large number of lens assemblies, detectors, and associated analog channels. One design being explored is to use fiber optics to bring incoming light into a single bundle that can then be transmitted onto a single CCD or FPA.
A wide range of sensors and optics exist that can be implemented into a SDS. Range, bandwidth, power, and cost become the ultimate driving factors for SDS systems. Low-cost thermopile detectors only offer a short range (~100m) and relatively slow sampling rates (10- 100 Hz). Higher-end Mercury Cadmium Telluride (HgCdTe) offer higher speed and higher detectivity, but at a high cost and with high power requirements for thermo-electrical cooling. For longer range detection associated with missile detection IRST systems, a range of options exists depending on cost and available power.
Heat Signature Determination
There exists a wide range of filters that can be used in combination with single- or multi-element broadband IR detectors. These filters allow for both broadband filtering and narrowband filtering. By implementing particular filters and multi-element detectors, the spherical sensor system can distinguish between different temperature objects. One common example is differentiating between a human IR signature and a ground vehicle IR signature. One of the detector elements can be fitted with a filter in the 3 - 5-um band, while another element will have an 8 - 12-um filter. In this scenario, the human IR signature will not produce a measurable response on the 3 - 5-um band, while the 8 - 12 band will produce a significant response. The vehicle temperature will generally overwhelm the 8 - 12-um sensor, but will also be present in the 3 - 5-um band. This technique can be applied to various spectral bands depending on application. Sensors with 10 channels or more will offer detailed multi-spectral differentiation.
The electrical subsystem, in combination with the algorithms, represents the heart of the SDS. The biggest hurdle is sampling 500+ sensors simultaneously and processing the data on an FPGA and DSP. Initial design is to obtain the 500 analog channels on the FPGA and then send data to a PC over an Ethernet connection for processing. An ADL945 duo core PC-104 is used for initial processing of the data. The PC-104 processing option enables a configurable system with the ability to add cards for auxiliary system control, but at the cost of higher power consumption. The PC-104 systems are ideal for integration on land- or air-based vehicles. As the algorithms develop, an additional DSP will be implemented for onboard high-speed processing. Basic calculations, such as finding peak maximums, can be accomplished on the DSP at very high speeds. The onboard DSP for data processing allows the SDS to operate at high speeds with low power consumption.
IRST systems classically employ distributed passive electro-optical systems to achieve a large FOR providing improved situational awareness. They are mainly used for detection, classification, and identification of targets within a line of sight. Their advantages over similar active technologies, such as radar, include low power consumption, high-speed scanning, high angular accuracy, high immunity to countermeasures, accurate target discrimination, and passive operation.
Since the 1960s, IRST systems have been used on military jet fighters and in the 1970s on naval ships. Advances in sensing have expanded the application of IRST technologies to a host of platforms for both defense and security operations, including marine vessels, aircraft, ground vehicles, man-portable units, and stationary mounts. IRST systems are now used to detect a multitude of targets, each with their distinct infrared signatures. Target types include small arms fire, missiles, RPGs, vehicles, and, of course, people. Optical ranges and IR band selection depend on the platform and application.
The SDS acts as the primary component of the IRST for immediate detection of threats. Once the SDS determines the angular coordinates and range of targets, the information is communicated to cameras for further identification. The SDS will continuously track targets and provide information to the cameras or other countermeasures, depending on the application.
The SSC’s primary strength lies in its continuous, high-speed, 360° horizontal by 360° vertical field of regard for automated multi-target detection and tracking in 3D. Due to its passive sensing, the unit observes its environment covertly while keeping power consumption and mechanical breakdown to a minimum.
The SSC’s individual sensors are tuned to fit the application by adjusting the optics, sensor type, and filtering. Spherical detector size generally increases for long-range, high-resolution systems. Installing additional SSC units to a platform, such as on large vessels, borders, or fence-lines, improves accuracy and coverage, and provides increased range-finding (triangulation) capabilities.
The SSC core technology offers effective high- and low-cost IRST systems with comprehensive benefits. Utilizing the real-time 3D angular data produced by the SSC, defense and security operations are enhanced for threat detection, reconnaissance, collision avoidance, intrusion detection, and search and rescue.
This article was written by Ryan Riel, Adam Calihman, David Thomson, Nicholas Jentzsch, and Matthew Eames of Lucid Dimensions, Louisville, CO. For more information, click here .