Predictive Maintenance
How it works:
AI models on devices analyze sensor data (vibration, temperature, etc.) from machines to detect early signs of wear or failure.Benefits:
Reduces unplanned downtime, extends equipment life
Visual Inspection
How it works:
Cameras combined with AI on device inspect products for defects in real time.Benefits:
Increases product quality, reduces scrap rates
Robotics/Cobots
How it works:
Physical AI powers real-time decision-making in robots, allowing them to adapt to changing environments and interact safely with humans.Benefits:
Increases efficiency, enhances flexibility, improves safety
Inventory Monitoring
How it works:
AI-enabled sensors and cameras track materials and parts, providing instant feedback on inventory levels.Benefits:
Minimizes shortages or overstocking
Anomaly Detection
How it works:
Physical AI models continuously monitor sensor data (e.g., pressure, current, temperature, acoustic signals) to detect patterns that deviate from normal operating conditions.Benefits:
Identifies faults, deviations in real-time, for immediate response
Measurement Analysis
How it works:
Physical AI processes data from vision systems, laser scanners, or coordinate measuring machines (CMMs) to assess part dimensions and tolerances during production.Benefits:
100% in-line inspection and auto correction without slowing production.
Assembly
Assembly Procedure
Equipment Operation Procedure
Safety & Training Procedure
Product Assembly Verification
- Hardware: MLSoC and MLSoC Modalix
- Software: Palette and Palette Edgematic
- Enable customer pipeline on SiMa.ai chip
- Integrate across multiple product lines and factories
- MLOPs
- Multimodal compute
- Industry leading frames-per-second performance
- Compute at the edge, enabling low latency, critical accuracy
Challenges
- Latency
- Throughput
- Varying video processing streams
Solution
- MLSoC – Multimodal compute at the edge
- Palette SDK & Edgematic – Deploy any model
Solution
- 10X improvement over incumbent solution
- Early defect detection
- Ease of integration in production
“AI has high strategic relevance for TRUMPF. Our expertise in laser processes and manufacturing helps us to develop smart software for production purposes. SiMa.ai is the ideal partner for this next big step towards intelligent industrial solutions”
Richard Bannmüller
CTO, TRUMPF Laser Technology
- Hardware: MLSoC and MLSoC Modalix
- Software: Palette and Palette Edgematic
- Develop perception and object detection stack
- Enable customer pipeline on SiMa.ai chip
- Deploy across multiple platforms
- 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