Xiaomi SU7 Boosts AEB, Fixes Steering Detection
💡SU7's AEB object expansion + torque detection key for automotive CV/ADAS builders.
⚡ 30-Second TL;DR
What Changed
AEB now detects vehicles, pedestrians, two-wheelers, barriers, barrels, boxes, stones, tires at 20-135km/h forward
Why It Matters
Enhances edge-case safety but exposes ADAS hardware trade-offs, influencing EV design choices.
What To Do Next
Test torque and expanded object detection in your ADAS computer vision pipelines.
Key Points
- •AEB now detects vehicles, pedestrians, two-wheelers, barriers, barrels, boxes, stones, tires at 20-135km/h forward
- •Speed coverage: front 1-135km/h, rear 1-30km/h; integrates with AES and MAI
- •Carbon fiber wheel needs 3-9 grip or torque sensor for hands-on detection, risks false alarms
- •Recommends enabling via Settings > Assisted Driving
🧠 Deep Insight
AI-generated analysis for this event — not the original article.
🔑 Enhanced Key Takeaways
- •The AEB update is part of Xiaomi's broader 'Xiaomi Pilot' software push, which leverages the NVIDIA Orin-X platform to process increased sensor fusion data for non-standard obstacle recognition.
- •The steering wheel detection issue stems from the physical material properties of the optional carbon fiber trim, which interferes with the capacitive touch sensors standard in the base model's leather-wrapped steering wheel.
- •Xiaomi has implemented a 'driver monitoring system' (DMS) camera-based fallback to verify driver attentiveness when the steering wheel torque sensor provides ambiguous data, reducing the frequency of false-positive warnings.
📊 Competitor Analysis▸ Show
| Feature | Xiaomi SU7 (Updated) | Tesla Model 3 (HW4) | XPeng P7i |
|---|---|---|---|
| AEB Obstacle Range | High (incl. debris) | High (Vision-only) | High (LiDAR-assisted) |
| Steering Detection | Torque/Capacitive Hybrid | Vision/Torque Hybrid | Capacitive |
| Pricing (CNY) | 215,900+ | 231,900+ | 223,900+ |
🛠️ Technical Deep Dive
- AEB System Architecture: Utilizes a multi-sensor fusion approach combining 128-line LiDAR, millimeter-wave radar, and 8MP high-definition cameras.
- Obstacle Recognition Logic: Employs a deep learning-based 'General Obstacle Detection' (GOD) network capable of identifying irregular shapes (e.g., traffic cones, fallen tires) not present in standard training datasets.
- Steering Detection Mechanism: The system utilizes a dual-input validation method: a capacitive sensor layer (in standard wheels) and a steering column torque sensor (for all wheels) to detect micro-adjustments in steering input.
🔮 Future ImplicationsAI analysis grounded in cited sources
⏳ Timeline
Weekly AI Recap
Read this week's curated digest of top AI events →
👉Related Updates
AI-curated news aggregator. All content rights belong to original publishers.
Original source: IT之家 ↗
This is a summary, not the original. Read the source, or get the weekly briefing.
Weekly AI briefing
One email a week. Unsubscribe anytime.