Bengaluru-based Entropik Technologies is focussing on offering safer mobility solutions using proprietary technology in the field of emotion research. The company tracks emotions at scale using technologies developed in-house. The core premise has been to bring more traction to existing man machine systems making experiences more seamless, safer and reliable. Auto Tech Review caught up with Ranjan Kumar, Founder & CEO, Entropik Technologies Pvt Ltd, to understand about its offerings.
Established in 2016, Entropik provides its customers with the ability to track their cognitive and emotional response at the point of origin, using new technologies such as brainwave mapping, facial coding and eye tracking. The company is also working on launching technologies like Object Detection and Vision AI that will play a revolutionary role in enhancing automotive road safety with a claimed accuracy of 15 % higher than any other systems available in the market. Entropik has received a funding of $ 1.1 mn during a pre-series A round, led by Bharat Innovation Fund and IDFC Parampara Capital.
EMOTION AI & ROAD SAFETY
Looking into the domain of automotive systems, Entropik ventured into the vertical with a problem statement of the mass losses to human life and capital caused due to poor road safety. Various studies released by National Highway Traffic Safety Administration (NHTSA) have revealed that around 1.3 mn people die in road accidents every year, accounting for 3,000 deaths a day and roughly 50 mn people are injured or disabled. The company worked on solving this problem statement using its Emotion Artificial Intelligence (AI) software and other proprietary technologies.
Relevant studies have showcased that 94 % accidents occur due to fatigue or stress-induced driving. With low driver alertness levels being the main cause of accidents, Entropik was able to insert their Emotion AI technology to comprehend and avert potential road accidents. Using just a simple camera, the company is able to track data points like distraction level, drowsiness level, yawning, eye closes and relevant stress markers in real time. Purely from a data statistics perspective, when the system is able to detect an action prior to an accident, it triggers a voice alert that may act as deterrent to accidents, he noted.
At present, OEMs track consumer behaviour using acceleration and driving pattern data pointers with multiple data sets available in the industry. He said it can also offer OEMs with the ability to track what a consumer is thinking and how they are behaving according to multiple stimuli, while experiencing a car. Using technologies such as brainwave mapping, etc., the company is able to offer OEMs and their product design teams data sets with relevant consumer feedback.
Kumar said the same emotion technology can also be used to optimise product experiences for customers. With rapid evolution in the sphere of mobility, technologies such as emotion tracking can also stimulate and change the environment inside a vehicle, using ambient lighting technologies available inside the car and based on the driver’s mood. Other areas of possible future application include seamless switching between manual drive and autonomous drive for vehicles equipped with relevant technologies.
CURRENT & FUTURE TECHNOLOGIES
Entropik focusses on three technologies – brainwave mapping, eye tracking and facial coding. For brainwave mapping, the company is using an EEG headset, which can be bought off the shelf and linked up to its Affect Lab brainwave mapping software. The EEG headset uses sensors that read neural impulses and the way in which the brain is firing. Kumar said the company has developed the software to read raw brain wave data according to a user’s thought process and then derive different data points like mental occurred stress, alertness level and emotional spectrum. The derivation from raw brainwave data to developing a final set of cognitive and emotional experiences is where the company’s core IQ lies, he pointed out.
Entropik has developed both facial coding and eye tracking software that derives data points using a standard camera, also available in some high-end luxury cars. Using a frontal camera, the software is able to track and evaluate facial expressions and derive emotions that are being exhibited at a micro level. Kumar said eye tracking is also executed using the same camera set-up and can track drowsiness, eye closures, distraction levels, etc.
The company has been able to induce machine learning capabilities for its software. This has been conducted in two parts – the first, in which two million data sets were created with multiple markers for facial expressions across different countries, cultures, religions and demographics. The second involves updating the database with new facial coding data, which whenever collected goes into a process of auto learning. The company upgrades its AI system every two weeks, Kumar said.
OBJECT DETECTION & VISION AI SYSTEMS
Entropik will also be launching its proprietary Object Detection and Vision AI systems in future. The derivation of data points will be captured through the same camera set-up used for its other systems, as it is largely scalable and widely available across geographies with a limited amount of friction in usage. Kumar said object detection is primarily focussed on automotive usage and how it can aid in tracking road conditions and situations by identifying objects around the vehicle. The company has also trained the system to use multiple data sets and will progressively make it a lot more exhaustive. Vision AI combines facial coding and object detection data sets and learns on its own. In totality, Vision AI analyses both data sets to predict the possibility of an accident by creating semantic meaning to each data point.
TEXT: Joshua David Luther