Imagimob
Real-time AI on the Edge

As the world continues to accelerate towards an autonomous future, it’s time to re-imagine the safety parameters pertaining to self-driving cars. Given today’s advanced driverless technologies, imagine being in an autonomous vehicle that is on the verge of crashing. In such scenarios, the traditional cloudcomputing systems fail to react instantly while sending data to the cloud for analysis, decision, and corresponding action. Alongside, to cope with a massive volume of unstructured data obtained from billions of connected IoT devices, as well as the network speed required to stream those remains a concern. Sweden-based Imagimob brings an advanced AI system with capabilities to deal with such situations. “We offer an AI system where the intelligence is executed on the device itself–on the edge,” states Anders Hardebring, CEO and Co-founder of Imagimob. Offering ‘Real-time AI on the edge’, Imagimob brings the core of data processing and networking closer to the connected devices. The company helps businesses improve safety, efficiency, and user experience in their products and services by enabling actionable insights on small devices.

The firm has developed and patented a software system for real-time AI on the edge–SensorBeat. It intakes sensor data or signals as input and translates them into useful information. Since this robust system is positioned close to the origin of data, SensorBeat is capable of creating a system with zero or very low latency between data creation and intelligent decision making. The AI system, being installed on the hardware device, is capable of collecting, monitoring, and analyzing data for itself, independent of internet connectivity.

Unlike most AI systems, SensorBeat AI is designed for easy and fast learning by leveraging even minor datasets. In order to lower the expenses for IoT devices, SensorBeat has been built in a way that it can run on low-cost microprocessors such as ARM Cortex M0. Since the AI system resides within the devices, it diminishes a major portion of power consumption and data communication costs in comparison to transmitting sensor data over Bluetooth LE or any other communication method.

SensorBeat works with all communication technologies, likely Bluetooth LE, 2G/3G/4G/5G, and WiFi. Using SensorBeat together with Low Power Wide Area Networks (LPWAN), such as NB-IoT, Sigfox or LORA is an ideal fit, since they share the same qualities that are important for IoT networks, low power and low cost.

Pushing the envelope of Motion AI, Imagimob has designed SensorBeat Motion—a software suite for Motion AI on the edge. The suite uses inputs from motion sensors such as accelerometers, gyroscopes, and magnetometers, and translates that data into actionable information following a much structured approach. From unlocking new dimensions for fitness trackers to aiding manufacturers with predictive maintenance technology–Imagimob’s unprecedented AI offerings have substantially enhanced its clients’ solutions portfolio, gearing them up for the future. The company has partnered with smartwatch manufacturer BellPal to implement SensorBeat into the watch’s hardware components which are for the elderly to understand their behavioural patterns and detect falls. Having checked globally for fall-detection algorithms BellPal decided upon Imagimob based on their expertise in Motion AI. While training Imagimob’s SensorBeat to recognize motion irregularities, BellPal simulated multiple fall scenarios by inviting healthy seniors to fall on mattresses while wearing Imagimob’s data capture device. The sensor data from different falls were then analysed to create various AI-patterns which can detect different types of falls. As SensorBeat became proficient enough to alert families or caregivers in case of serious falls, the software was configured and implemented into BellPal’s smartwatch.

Apart from BellPal, Imagimob’s astounding Motion AI technology has been leveraged by enterprises from multiple verticals to deliver unprecedented user experiences to their customers. As the upcoming generations of applications are getting well poised to be pillared on ‘real-time AI on the edge’, Imagimob is constantly striving to push the boundaries of this technology further. “We want to continue being the best at what we’re solely doing,” Hardebring concludes.