In an interview, Anoop Aggarwal, Head of Automotive Sales at Analog Devices, spoke with AutoEV Times about the shift from advanced ADAS to Level 3 autonomy in India. He highlighted how software-defined E/E architectures, sensor fusion, and edge intelligence are reshaping vehicle safety, driver experience, and real-time perception. The discussion also explored scalable, cost-effective solutions, cloud–edge integration, and ADI’s role in accelerating interoperable, autonomous vehicle technologies.
Read the full interview here:
AET: As the industry moves from advanced ADAS and Level 2 systems toward Level 3 autonomy by 2026, what fundamental shifts are required in vehicle safety architecture and driver experience—particularly for real-world Indian driving conditions?
Anoop: As India’s automotive industry moves steadily up the autonomy curve, the shift from advanced ADAS and Level 2 systems to Level 3 autonomy marks far more than an incremental step forward. For a modern India, this transition demands a holistic re-architecture of the vehicle’s safety system, especially when contextualised against real-world operating conditions. Successful Level 3 deployment will require redundant sensor pathways, deterministic fail-safe compute, and tightly synchronised perception stacks that can manage chaotic, unstructured environments seen on Indian roads. Indian automotive adoption data from 2025 underscores robust growth in Level 2 penetration, rising adoption from ~6.2 % to ~8.3 % in H1 2025, driven by SUVs and EV platforms, a clear signal that systems are maturing but remain mostly “driver-in-the-loop” today.
Fundamentally, automotive safety architecture must evolve toward a software-defined, zonal E/E configuration that supports modular safety domains, cross-domain checks and digital threads from sensing through actuation. This is where open standards and uniform data transport, such as the emerging OpenGMSL initiative, become critical, enabling consistent, low-latency data flows across multiple tiers of safety functions.
AET: High-performance ECUs and edge processing are becoming central to autonomous decision-making. How do you see edge intelligence reshaping real-time perception, latency management, and system reliability in autonomous and connected vehicles?
Anoop: One of the biggest shifts we’re seeing in autonomy right now is where the intelligence actually sits in the vehicle. High-performance ECUs and edge processing will be central to autonomy’s next chapter because they eliminate the latency, availability gaps and cost exposure of cloud-centric models. Running inference and multi-sensor fusion directly on-vehicle allows deterministic reaction timing, critical where milliseconds can influence safety decisions and reduces reliance on variable network conditions. In practical engineering terms, scaling compute close to sensor clusters, enabled by architectures such as zonal domain controllers combined with high-bandwidth interconnects reduces system complexity and preserves thermal, power and reliability margins even in constrained automotive environments.
AET: AI-driven mobility increasingly relies on seamless cloud–edge synergy. From your perspective, how can this integration enhance predictive safety, vehicle efficiency, and over-the-air intelligence without compromising data security or latency?
Anoop: When we talk about the cloud–edge continuum in mobility, it’s really about being deliberate with where compute delivers the most value. Real-time inference and safety-critical decisions have to sit at the edge, on the vehicle, while the cloud plays a complementary role in large-scale model training, fleet analytics and over-the-air intelligence. When this balance is executed well, predictive safety improves meaningfully, models can be trained and validated in the cloud, then securely deployed back into edge ECUs without compromising latency or system integrity.Edge systems help ensure that critical safety decisions remain local and deterministic, while cloud platforms enable predictive diagnostics, usage-based update roll-outs and fleet performance optimisation. From a security standpoint, robust cryptographic anchoring and secure boot mechanisms are necessary to ensure OTA updates and data exchanges are resilient against threats, something automotive OEMs cannot trade off in pursuit of latency advantages. While this detailed cloud–edge interplay is not fully unfolded in 2025, the trajectory captured in connected-tech discourses suggests a pragmatic balance between cloud agility and edge determinism.
AET: Sensor fusion is critical to building a holistic perception layer. How do radar, LiDAR, cameras, and IMUs complement each other in enabling Level 3 autonomy, and what engineering challenges must be addressed to ensure robustness in complex environments?
Anoop: Level 3 autonomy fundamentally relies on sensor fusion, as no single sensing modality can deliver reliable perception on its own. Cameras provide detailed visual context, radar offers robust range and velocity measurements under challenging conditions, LiDAR delivers precise 3D spatial mapping, and IMUs supply critical motion and orientation information. Ensuring that all sensor data is accurately synchronised and of high integrity is essential for AI systems to make consistent real‑time decisions, particularly in complex and unstructured environments such as Indian roads. This integration of complementary sensors forms the foundation for transitioning from driver assistance towards autonomous vehicle operation.
Beyond sensor complementarity, the key challenge for Level 3 autonomy lies in reliably fusing heterogeneous data in real-world conditions. As highlighted in recent sensor-fusion research, achieving precise spatio-temporal alignment across cameras, radar, LiDAR and IMUs is critical; even minor timing or calibration errors can destabilise object perception in dense, dynamic traffic typical of Indian roads. Fusion systems must also remain resilient to modality degradation caused by rain, dust, occlusion or poor lighting, while operating under strict on-vehicle compute and latency constraints. Addressing these challenges through robust synchronisation, failure detection and computationally efficient fusion architectures is essential to delivering consistent, safety-grade perception in complex Indian driving environments.
AET: Cost efficiency remains a key factor for OEMs and Tier-1 suppliers, especially in emerging markets like India. What growth opportunities do you see for scalable, cost-optimized ADAS and autonomy solutions without diluting safety or performance?
Anoop: At ADI, we see significant growth opportunities by delivering scalable, modular ADAS and autonomy solutions that optimise system-level design without compromising safety or performance. Our GMSL™ technology, for example, consolidates video, data, power, and control over a single interface, lowering complexity and cost while maintaining the high-bandwidth, low-latency performance that advanced ADAS demands. Open-standard adoption further expands supplier choice and accelerates development, enabling local ecosystems to grow efficiently. By combining sensor fusion and software-defined architectures, compute and sensing resources can be shared across multiple ADAS and autonomy functions, making solutions adaptable and scalable. In India, this approach allows OEMs to manage cost pressures while meeting diverse environmental and road conditions, supporting faster adoption of safe, high-performance autonomous technologies.
AET: Analog Devices is playing a significant role in advanced sensing, AI-enabled solutions, and initiatives like OpenGMSL. How does ADI’s approach to standardised, interoperable in-vehicle connectivity accelerate the transition to safer, software-defined, and autonomous vehicles?
Anoop: At ADI, we see standardised and interoperable in-vehicle connectivity as a key accelerator for safer, software-defined, and autonomous vehicles. Through the OpenGMSL™ Association, which we champion alongside OEMs, Tier-1 suppliers, and technology partners, we are establishing a global open standard for high-speed video and data transmission in vehicles. Building on our GMSL™ technology, which has already achieved broad adoption, OpenGMSL ensures that multivendor products work seamlessly together, from sensors and cameras to displays and compute modules. By creating a consistent, compliant framework, we reduce integration complexity, accelerate development timelines, and support innovation across ADAS, autonomy, and infotainment. Ultimately, this approach helps the entire automotive ecosystem move faster toward safe, software-defined, and autonomous vehicles.
