Edge AI in the Driver’s Seat: How Elektrobit Champions Real-Time, Privacy-First Emotion Detection for Automotive Innovation

#EdgeAI #AutomotiveInnovation #InCabinExperience #EmotionDetection #ADAS #Elektrobit

October 2025 : Shivaranjini , Senior Technical Architect, Elektrobit India , in an exclusive interaction with our Manish Kulkarni , shared insights on edge-native AI redefining in-cabin safety and personalization by enabling real-time emotion detection with uncompromising privacy for the next generation of automotive experiences.

1. How is Elektrobit using edge AI to achieve real-time, privacy-friendly emotion detection in vehicles?
Elektrobit has the capabilities and expertise to develop end-to-end solutions for in-cabin monitoring systems, including emotion detection in vehicles. Our priority is building edge native AI models, with privacy embedded at every stage of execution. Our approach combines optimized on-device models with very strict data governance, so we deliver real-time emotion awareness without exporting raw biometrics off the vehicle.
Key elements include:
● Edge-first model architecture : compact, latency-optimized models that run locally in in-vehicle for instant response and no dependency on cloud connectivity.
● on device processing: raw sensor inputs (camera, audio, etc) are transformed into abstract features locally, with no storage or transmission of identifiable imagery or audio to external services like cloud, for example.
● Customization of data and execution policies: Data storage, retention, and model execution are fully customizable, which can be aligned as per the privacy policy.

2. Where does Elektrobit see the greatest impact of its technology in ADAS, the in-cabin experience, and generative AI?
ADAS:
● We have seen the greatest impact in ADAS through AI-driven perception and decision systems. Activities like object, obstacle, lanes, and hazards detection can be achieved with higher fidelity, enabling advanced safety features like emergency braking, lane-keeping, and parking assistant systems.
● Generative AI helps in the generation of synthetic data and training environment creation, which helps in accelerating technology development and ensuring robustness across various scenarios.
In cabin :
● Elektrobit Next Generation Cockpit is a great example of shaping the in-cabin experience. In combination with AI, the cockpit solution provides a rich display, a customizable dashboard, and dynamic selection and adaptation of the theme. Designed with flexibility in mind, it supports continuous enhancements and integrations. Elektrobit has established partnerships with leading technology players like Google, enabling us to jointly leverage and integrate innovations such as Google Gemini for enhancing in-cabin experiences.
● In-cabin experiences today are enriched with driver and occupant monitoring system algorithms. These technologies are mainly driven by AI, which senses driver alertness, mood, and activity of the occupants. Safety is enabled by giving timely alerts and escalating in situations like the case of driver drowsiness/fatigue. The technology also helps in further personalization of the cabin, like adaptive seat adjustment, media selection, etc. Clubbed with Generative AI voice-based assistants, users enjoy natural interaction with the vehicle.
Generative AI:
● Generative AI helps various areas, like in the synthesis of data for various ADAS requirements, automation of UI/UX content generation for in-cabin systems, and voice-based interaction with the vehicle, etc. And, on the engineering side, it helps in development, coding, debugging, and test automation activities.

3. Can you share examples where emotion data has directly improved ADAS performance or driver support?
The overall idea here is to build a pipeline around the emotion detection activity that detects the distraction, fatigue, frustration, and overall mood of the driver. The pipeline activities can vary from analysing the driving pattern based on the emotions of the driver to fusing the outcome of emotion detection with the ADAS system for better decision-making.
ADAS behaviour can be tuned in combination with the emotion detected output:
● When the driver is under stress: reducing less critical alerts, adopting smooth driving behaviour, switching to a calmer voice in voice-based assistants, etc.
● When the driver is fatigued: trigger lane keeping interventions, increase alert frequency, recommend rest stops, etc.

4. How has the in-cabin experience evolved, and what is Elektrobit’s vision for its future?
Today’s In-cabin experience has evolved into a more connected, intelligent environment, with interior sensing, immersive displays, and integration of voice-based assistants. An earlier static environment has turned into a more experience-rich environment focused on comfort and safety.
Elektrobit envisions the next generation of in-cabin systems as an intelligent, adaptive, and highly personalized environment . Our Next-Gen Digital Cockpit will integrate emotion and activity detection, multi-modal interaction (voice, gesture, and touch), and context-driven personalization to adjust lighting, climate, and media. We aim for the creation of human human-centric system that understands every journey and every occupant.

5. Could you shed light on the ways EB robinos enable more intuitive, emotion-aware voice assistance?
EB robinos Predictor is a product that provides a hardware-agnostic electronic horizon for accurate and up-to-date information about the road ahead for predictive ADAS and automated driving. Vehicles and drivers greatly benefit from information beyond the sensor horizon for a safer and more comfortable ride.
The product has 2 modules:
● Provider module: for providing eHorizon data in ADASIS v2/v3 protocol format with the help of underlying SD/HD map data)
● Reconstructor module: reassemble and efficiently store relevant electronic horizon data
Emotion-aware voice assistance feature is NOT part of the core product offering. Further pipeline applications can consume this data and deliver voice-based interactions depending on the emotions of the driver. For example, if a hairpin curve is ahead on the road and the driver is fatigued, the application can deliver voice alerts.
What safeguards does Elektrobit employ to balance emotion detection capabilities with user privacy?
At Elektrobit, we design emotion detection based on the key principles of privacy and transparency. Our AI models are optimized for edge-device processing, ensuring all the biometric data is processed locally within the vehicle. This approach minimizes transmission to external servers. In cases when the data must be transmitted to external servers, it is anonymized and encrypted before transmission. Further, only the extracted features are transmitted rather than the complete biometric data, safeguarding the user's privacy.

6. How is the adoption of edge AI, ADAS, and emotion-aware technologies evolving in the Indian automotive market, and what challenges or opportunities do you see ahead?
In the Indian automotive market, OEMs and Tier-1 suppliers are steadily increasing in investments in technologies like Software Driven Vehicle (SDV), AI, cloud, and ADAS technologies. This growth is also fuelled by falling cost of sensor prices and compute environments, advancements in AI, a growing pool of skilled experts, and rising consumer expectations for both safety and comfort.
Key challenges include variable road infrastructure, cost sensitivity in a consumer-driven market, and heightened privacy concerns around biometric and behavioural data. At the same time, significant opportunities include the rollout of 5G network, maturation of AI ecosystem and cloud infrastructure, and the shift towards SDV platforms.

Shivaranjini, Senior Technical Architect at Elektrobit India, is a seasoned technology leader with over 21 years of experience across automotive systems, AI, and cloud solutions. With an M.S. in Software Systems and a B.E. in Computer Science, she has led end-to-end development of ADAS, DMS, computer vision, and AI/ML-based innovations at Bharat Electronics, Gnostice, and Elektrobit.

Pro MFG Media Newsletter

Get regular manufacturing insights that matter and enable business.