ISTI brings together several experienced researchers with Ph.D. degrees whose expertise encompasses such areas as signal and image processing, distributed sensor networks, communications and data compression, detection, classification and tracking, neural networks and adaptive systems.

ISTI has been involved in research and development in the areas of unattended acoustic sensors for detection, localization, classification, and tracking of militarily significant acoustic sources. Examples of such sources include: gunshots, RPG and artillery fires, and ground or low-flying vehicles. Under the SBIR funding, ISTI has developed breakthrough system-level hardware and algorithm solutions to these high priority problems using distributed wireless sensors. ISTI has recently developed and prototyped low-cost, small-size and stealthy sensor nodes (click products for more information) that offer unique benefits. These benefits include: wide coverage area, robustness to acoustic multi-path effects in urban terrain, ability to capture and process both muzzle blast and ballistic shockwave for gunshots, ability to detect and localize gunshots as well as other sources e.g. ground or airborne vehicles, self-localization ability, potential for fixed site deployment or portable (e.g. individuals forming a mobile sensor network). To overcome the bandwidth limitations in wireless sensor networks; sensor-level data compression systems are being developed that offer great bit rate reduction without sacrificing the essential features of the targets. This allows for large scale proliferation of these low-cost sensor nodes in a surveillance area. Algorithms are developed for accurate detection and localization of gunshots from muzzle blast and/or shockwave. The overall system is being designed to provide the users with a single integrated acoustic event assessment picture based upon the history of occurrence and types of events detected, localized and classified by the system.

ISTI has also been involved in developing adaptable sonar image classification and retrieval system that can learn in-situ in new environmental and operating conditions. This technology can be applied to many applications that require in-situ learning of classification and retrieval systems.