Acoustic Event Detection

Solwit AED is an Edge Machine Learning neural network that offers a breakthrough technology to identify and analyze sounds within residential, city, automotive and factory environments.

AED detects audio events and acoustic scenes by analyzing audio content in real time

AED aims at processing a continuous acoustic signal and converting it to a pre-programmed behavior of the system – triggering a corresponding custom response.


When an acoustic event occurs, AED detects the sound and recognizes its character.


The system is trained to recognize a large number of acoustic events and scenes, relevant to the customers' needs.


The audio event is reported by customizable means (i.e. sending information to a smart device, calling for the emergency, starting alarms, etc.)


The AED firmware allows for fast porting to various operating systems and hardware platforms. Starting from Linux and multicore ARM solutions, through single core microcontrollers, RTOS or bare metal platforms, and ending with FPGA systems.

The AED firmware is designed in an object-oriented architecture which allows us to trim our solution perfectly to the customer requirements.


The AED solution can work as a stand-alone system or be part of a customer’s solution. Its firmware allows to change system resources utilization to better fit the customer needs.

Additionally, it can support different numbers and types of sounds. The AED firmware is ready to support various technologies, such as beamforming, noise reduction, and microphone matrices.


The AED firmware and hardware development process meets very strict standards.

An amazing development team with background in research and development of audio systems for industries such as automotive, military, and transportation controls AED’s development and quality.

Additionally, a multilevel testing process assures that the customer gets the solution of the highest quality.

Possible use cases

AED can be utilized in a variety of applications, including context-based indexing and retrieval in multimedia databases, unobtrusive monitoring in health care, and surveillance.


  • non-invasive IoT Retrofit,
  • adaptable to any machine,
  • safety alarms, machine failures, explosions, etc.,
  • portable (battery).


  • detecting potential injuries by identifying falls, screams, and other hazards
  • detecting potential threats (break-ins),
  • identifying a specific sound and alerting the proper authorities, family, or residents.


  • as a stand-alone sensor or part of a system,
  • activates selected parts of a system or informs residents about a dangerous event,
  • supports deaf and old people in cases of danger.


  • as part of a system,
  • informs appropriate authorities about violence on streets or excessively loud vehicles,
  • safety alarms: machine failure / explosion / screaming.


We analyze, process, and teach our AED to recognize audio events and acoustic scenes in our dedicated Sound Labs.

Our engineers who specializes in the science of sound work on noise emitted by factories, industrial plantses, day-to-day everyday life environments or businesses facilities.

Our acoustical engineers develop the AED technology by using elevated high -quality, real-world data sounds which gives our solutions the ability to identify and analyze sounds for residential, city, automotive, and factory surroundings environments.