🏠Freshcollected in 2m

New Deepal S05 SUV Launches with Advanced AI Driving

New Deepal S05 SUV Launches with Advanced AI Driving
PostLinkedIn
🏠Read original on IT之家

💡See how end-to-end AI and 3nm chips are transforming the smart cockpit and autonomous driving experience.

⚡ 30-Second TL;DR

What Changed

Equipped with 1 LiDAR and 26 sensors for full-domain perception.

Why It Matters

The integration of end-to-end AI algorithms and high-compute chips in mass-market SUVs demonstrates the rapid democratization of high-level autonomous driving features.

What To Do Next

Evaluate the performance of the Horizon J6M chip in handling end-to-end autonomous driving tasks compared to global competitors.

Who should care:Developers & AI Engineers

Key Points

  • Equipped with 1 LiDAR and 26 sensors for full-domain perception.
  • Powered by dual Horizon J6M chips for high-level autonomous driving.
  • Features DEEPAL OS 4.2 with large-model integrated voice interaction.
  • Includes 3nm cockpit chip for improved multi-tasking and OTA potential.

🧠 Deep Insight

AI-generated analysis for this event.

🔑 Enhanced Key Takeaways

  • The Deepal S05 is positioned as a compact SUV targeting younger demographics, emphasizing a 'smart' design language that includes an optional external projection system capable of 120-inch displays.
  • The vehicle utilizes a 'Skyeye' system architecture that integrates the Horizon J6M chips with a high-performance perception stack to enable urban NOA (Navigate on Autopilot) capabilities.
  • Deepal has implemented a 'Deepal AD' intelligent driving platform which supports parking assistance features including remote parking and narrow-space maneuvering.
  • The 3nm cockpit chip mentioned is widely identified in industry reports as the Qualcomm Snapdragon 8295, which provides the necessary TOPS for the vehicle's 3D HMI and complex voice processing.
  • The S05 series offers both pure electric (BEV) and extended-range (EREV) powertrain options, reflecting Deepal's dual-track strategy to maximize market penetration.
📊 Competitor Analysis▸ Show
FeatureDeepal S05BYD Yuan Plus (Atto 3)XPeng Mona M03
Intelligent DrivingHorizon J6M (Urban NOA)L2 ADASAdvanced XNGP
Cockpit ChipSnapdragon 8295 (3nm)Snapdragon 888Snapdragon 8155
Primary MarketTech-focused YouthMass MarketAI-First Compact
PowertrainBEV / EREVBEVBEV

🛠️ Technical Deep Dive

  • Perception Stack: Employs a fusion of 1 LiDAR, 5 millimeter-wave radars, 12 ultrasonic radars, and 8 high-definition cameras.
  • Computing Power: The dual Horizon J6M configuration provides approximately 384 TOPS of AI computing power, optimized for end-to-end transformer-based models.
  • Architecture: Built on the EPA1 platform, which supports high-voltage fast charging and integrated thermal management for the battery pack.
  • Voice Interaction: The large-model integration supports multi-modal interaction, including continuous dialogue, zone-specific voice recognition, and offline command execution.

🔮 Future ImplicationsAI analysis grounded in cited sources

Deepal will transition its entire product lineup to the J6M/J6P chip architecture by 2027.
The successful integration of Horizon Robotics' latest chips in the S05 suggests a standardized hardware strategy to reduce software development costs across the brand.
The S05 will see a significant increase in urban NOA adoption rates compared to previous Deepal models.
The combination of 384 TOPS of computing power and end-to-end algorithms significantly lowers the technical barrier for complex city driving scenarios.

Timeline

2023-04
Deepal brand officially announces its independent development strategy at the Shanghai Auto Show.
2024-03
Deepal announces a strategic partnership with Horizon Robotics to utilize J6 series chips.
2024-08
Deepal S05 makes its public debut, highlighting the new intelligent driving and cockpit features.
2024-10
Official market launch of the Deepal S05 in China.
📰

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