ADAS (Advanced Driver Assistance Systems)
We Enable:
• Real-time perception (object detection, lane keeping, pedestrian tracking)• Sensor fusion (camera + LiDAR + radar)
• Adaptive cruise control and Driver monitoring systems (DMS)
Why SiMa:
Low-latency response, energy-efficient ML pipelines, and support for diverse sensor inputs
Autonomous Driving (L2+ to L5)
We Enable:
• Environmental perception and path planning and Real-time decision-making• Low-power inference across multiple neural networks
• Fail-safe perception for redundancy layers
Why SiMa:
Deterministic performance, tight control over model pipeline execution, and hardware/software co-optimization
In-Cabin AI & UX
We Enable:
• Occupant detection & classification• Gesture recognition, voice control, Emotion/intent detection
• Personalized infotainment systems
Why SiMa:
Enables smarter human-machine interfaces with high model accuracy at minimal power draw
Fleet & Telematics Intelligence
We Enable:
• Edge video analytics for vehicle health and surroundings• Event-based video logging (crash, intrusion, road condition)
• Real-time alerts without cloud latency, battery-conscious deployments
Why SiMa:
On-device processing reduces cloud costs and improves latency for mobile/fleet-based deployments
Predictive Maintenance/Smart Diagnostics
We Enable:
• Detect abnormal sounds/vibrations via ML• AI-based error code prediction and prioritization
• Real-time system monitoring and alerts
Why SiMa:
Efficient ML inference pipelines allow continuous diagnostics with minimal overhead
Smart Traffic & V2X Systems
We Enable:
• Edge processing for roadside units (RSUs)• Real-time traffic flow monitoring, incident detection and alert propagation
• Situational awareness for connected vehicles
Why SiMa:
High-performance edge compute in embedded environments ensures responsiveness and reliability
Driver & Vehicle Safety
Lane Keep Assist
Pedestrian & Cyclist Detection
Forward Collision Warning
Blind Spot Monitoring
Driver Drowsiness Detection
Parking Assistance
Autonomous Driving/Perception
Multi-modal Sensor Fusion (Camera + LiDAR + Radar)
Real-time Object Detection & Tracking
Traffic Sign & Signal Recognition
Road Edge and Lane Line Detection
Path Planning and Obstacle Avoidance
Low-latency Inference for Redundant Safety Systems
In-Cabin Monitoring
Child Presence, Seatbelt Detection
Gesture Recognition HMI Control
Face Recognition Personalization
Edge Voice Command Processing
Connected Vehicle Systems
Real-time Traffic Analysis
Predictive Maintenance
Fleet & Commercial Vehicles
Driver Behavior & Scoring
Autonomous Vehicles
Multiple models, camera and LIDAR integration
SiMa.ai Solution Detail
Products
- Hardware: MLSoC and MLSoC Modalix
- Software: Palette and Palette Edgematic
Services
- Develop perception and object detection stack
- Enable customer pipeline on SiMa.ai chip
- Deploy across multiple platforms
Solution Overview
- Multimodal compute enabling perception, mapping, localization, and navigation
- Industry leading frames-per-second performance
- Compute at the edge, enabling low latency
Challenges
- Multimodal low latency compute at the edge
- Sensor flexibility
- Low power profile for mobility platform
Solution
- Simultaneous Localization and Mapping (SLAM)
- 16 channel LIDAR input at 10 FPS
- Multi-camera input (6) – object detection at 30 FPS
Takeaways
- State of the art mobility platform that is safe
- Cost effective scaling at the edge
“AI is transforming robotics, and at Virya, we’re harnessing its potential to deliver real value. Our partnership with SiMa.ai brings powerful edge-AI capabilities for real-time, low-power decision-making—enabling smarter, faster, and scalable solutions. We’re proud to be among the first in India’s autonomous mobility space to integrate their platform.”
Saba Gurusubramanian
CTO, Virya Autonomous Technologies Pvt Ltd