IDENTIFICATION OF ACOUSTIC SIGNALS in noisy environments remains one of the most difficult of signal processing problems and is a major obstacle to the high degrees of accuracy and speed needed to identify suspicious sounds in high-security, high-safety environments. It has been argued that the brain is an existence proof of a near-optimal "model-based" system for pattern recognition: through evolution, the brain has found the optimal feature models to guide the search over an almost infinite input space such that efficient pattern recognition can be achieved, even in the presence of noise. Information in the brain is coded in terms of variation in the sequence of all-or-none temporal patterns, transmitted between nerve cells. Identifying the nonlinear input/output properties of neurons involved in forming memories for new patterns, and developing mathematical models of those nonlinear properties, can provide a revolutionary pathway to neural-based classification. More information on this research can be found at Theodore W. Berger Laboratories at USC, for which Safety Dynamics, Inc. serves as commercialization arm.

EFFECTIVE SITUATIONAL AWARENESS requires unambiguous identification of threats, and accurate localization of the source of those threats, so that optimal decisions can be made with respect to evasive actions or aggressive engagement of the enemy — decisions that have life or death consequences. Over the last decade, the military has connected nearly all its command posts and all its vehicles into a kind of internet for battle which allows them to, at the very least, see each battlespace component's location and better coordinate attacks. Individual warfighters, however, still remain largely off the grid. This is problematic because counter-insurgency fights are almost wholly dependent on small groups of Soldiers — who need technological advancements to keep them informed and aware.

DEPARTMENT OF DEFENSE-FUNDED RESEARCH for the development of smart sensors that can provide warfighters with information about the battlespace that will create situational awareness — the sensing, understanding, and effectively responding to events as they unfold in real-time, particularly in the context of safety in battle or en route to battle. Safety Dynamics is a collaboration between the pure research performed at the Berger Labs of USC and the implementation of prototypes that are biologically-based by Safety Dynamics; which come together to complete a product development life cycle that can serve both military and homeland security markets. Above and beyond serving contracts with the Army Research Office, the Office of Naval Research and the Department of Defense, there is a greater vision for the development of a library of sounds allowing customers to choose tailor-made acoustic recognition sound signatures with localization and intelligent video analysis solutions. The idea is to make the sensors smarter, make the surveillance obvious or obfuscated and make the warfighter safer.

THROUGH FUNDING FROM THE ARMY, a gunshot location system has been developed for urban crime prevention and for protection of military troops. DSNNs have been trained to recognize gunshots from .22, .25, .32, .38, .40, .45, .57, 9mm Caliber Pistols; .22, .223, .308 Caliber Rifles; .410, .12 gauge shotguns; AK-74 (5.45 x 39); SKS (7.62 x 39); Mosin Nagant (7.62 x 54R); RPK (7.62 x 54R); SVD (7.62 x 54R); Tokerev (7.62 x 25 mm); and Makarov (9 x18 mm) attack rifles, but not to respond to other loud noises, e.g., fireworks, truck and bus engines and motorcycle backfires. Classification rates are in the range of 90-95% for distances of 1000 feet or 304.8 meters, except for .22 caliber events, which are detected at that accuracy only to 350 feet or 106.68 meters. The DSNN technology was licensed by Safety Dynamics Inc., which released its first product, SENTRI, a gunshot location system designed to identify gunshot events in urban environments, to determine the location and origin of the gunshot, to direct a video camera to collect data on the gunshot source and to wirelessly send a signal to a central command where officers can immediately begin monitoring the video stream. A small number of other gunshot location systems are commercially available, but none offer the same functionality as the DSNN-based system. Because of the fast computational power of the DSNN classifiers, derived from neural networks, the determination of a gunshot occurs in about 4/10ths of a second.

OTHER SYSTEMS RELY on collecting the data from installed microphones (typically in a grid formation) and shipping that raw data to a central processing station along a network. Manufacturers of these types of systems claim a gunshot event can be confirmed in 5-12 seconds. Once there is a positive identification of a gunshot, knowing exactly where to move a camera or an optical device is paramount to identifying the threat for situational awareness. The DSNN-based gunshot system is commercially-proven precisely to locate the origin of a gunshot within ±1°. Assuming the distance between the shooter and the gunshot detector is 600 feet, a ±1° differential would mean that the sensor could be off by a maximum of 20.94 feet. Based on competing manufacturer’s claims of localization of gunshot within 25 meters, competing systems have the potential to be off by as much as 82.5 feet. A major advantage to the DSNN-based system is that it not only calculates the azimuth of the event (1-360°), but also the elevation of the event. For the case of a shooter firing down at a vehicle in an urban environment, the SENTRI system locates the building from which the shooter discharged and from which window ledge

Расшифровываю, если кто не понял.


SENTRI - система интеллектуальных нейроноподобных датчиков для первичного распознавания и идентификации опасности
В спорткомплексе "Олимпийский" прошла 11-я Московская международная выставка "Охрана, безопасность и противопожарная защита". Компания Innovative Crime Prevention, эксклюзивный представитель компании Safety Dynamics, впервые продемонстрировала систему SENTRI (Smart Sensor Enabled Neural Threat Recognition and Identification) для обеспечения превентивных мер по борьбе с криминогенной обстановкой на улицах городов и населенных пунктов.
Технология, легшая в основу системы, была разработана Теодором Бергером, директором отделения нейронного проектирования университета Южной Калифорнии и одного из соучредителей компании Safety Dynamics. За основу технологии Бергер взял гиппокамп человека - часть мозга, отвечающую за память, формирующую модели для каждого нового звука и анализирующую сохраненную в памяти информацию о шумах, позволяя их распознавать. Доктор Бергер "обучал" компьютер распознавать бесчисленные вариации звуков, производимых при стрельбе из огнестрельного оружия.
Принцип работы системы
В комплект системы при поставке входят: 4-входная акустическая плата SENTRI, источник питания переменного тока, комплект из 4 микрофонов.
Звуковой распознаватель выстрела из огнестрельного оружия состоит из специальной системы с четырьмя внешними микрофонами, поворотной купольной камеры непрерывного наблюдения с обзором в 360 град. и некоторых других компонентов. Система защищена пуленепробиваемым кожухом.
В момент выстрела или другого звукового сигнала (эффективный радиус ее действия - 2 городских квартала) система SENTRI распознает и определяет местоположение происшествия и в течение нескольких секунд посылает сигнал о событии в правоохранительные органы. Камера разворачивается в направлении источника звука, так что оператор может наблюдать происходящее с монитора, установленного в центре наблюдения или в автомобиле. Сигнал передается по мобильной связи.
Области применения системы SENTRI
Данные системы могут быть запрограммированы на распознавание любого звукового сигнала. Например, уже существуют системы для принятия превентивных мер по борьбе с нелегальной вырубкой леса. Такие устройства реагируют на звук цепной пилы, на шум дизельного грузового автомобиля.
Система может устанавливаться мобильно на платформе автомобиля и использоваться для временных наблюдений в местах, где проводятся многолюдные мероприятия (футбольные матчи, концерты и т.д.).
SENTRI может быть использована для защиты банков, инфраструктур, нефтегазовых проводов, атомных станций и других стратегических военных объектов.

скоро тбфорум кстати, поедете?

Пока под вопросом

Нет, у нас лады по дорогам страны ездят, а вы о таком.

сапожник не знает есть ли в стране сапоги

Это скорее не сапоги а сапоги-скороходы

Грядет тотальный контроль. Хорошо это или плохо? Вспоминаю книгу "1984".....

тебя посодют, а ты ен воруй