BYD Patents AI System to Detect Living Beings Under Parked Vehicles
BYD's latest patent isn't about range or charging speed — it's about not running over a child or a dog hiding under your car. The system uses baseline imaging to detect living beings beneath a parked vehicle before the driver pulls away.
Explanation
Most car safety systems kick in while you're moving. BYD is pushing the boundary earlier — to the moment before you even start rolling.
The newly unveiled system creates a "baseline image" of the underside of a parked vehicle when no one is around. When the driver returns and prepares to move, the system compares the current undercarriage view against that baseline. Any difference — a child, a pet, a person — triggers an alert before the car moves.
This matters because low-speed driveway and parking-lot incidents are a persistent blind spot in automotive safety. Rear-view cameras help, but they don't cover what's directly beneath the chassis. Small children and animals are especially vulnerable because they're below the sightline of standard sensors and cameras.
BYD has filed a patent on the approach, which signals intent to productionize rather than just research. Whether it ships as a standard feature or an optional add-on — and on which models — hasn't been confirmed.
The incremental nature of the signal is worth noting: this is a patent disclosure, not a product launch. The gap between a filed patent and a feature in a shipping vehicle can be years, or never. Still, the underlying problem is real, the baseline-imaging approach is technically straightforward, and BYD has the vertical integration to move fast if it chooses to.
BYD's patented system addresses a well-documented gap in automotive safety coverage: the under-vehicle blind zone during low-speed departure from a stationary position. Existing solutions — ultrasonic proximity sensors, 360° surround-view cameras, automatic emergency braking — are optimized for dynamic scenarios and typically have minimum detection distances that exclude objects flush against or beneath the chassis.
The core mechanism is a baseline-differencing approach. The system captures a reference image of the undercarriage environment when the vehicle is stationary and the area is confirmed clear, then performs a delta comparison at departure time. Anomalies above a threshold — consistent with a living being's thermal or visual signature — suppress movement or trigger a driver alert. The "living being" framing suggests the system may incorporate thermal or motion-based liveness detection rather than pure optical differencing, though the excerpt doesn't confirm the sensor modality.
From a systems perspective, this is architecturally similar to intrusion-detection logic used in security cameras, applied to a mobile platform with a hard real-time constraint. The novelty likely lies in the specific implementation: sensor placement, baseline refresh logic, environmental robustness (rain, shadows, uneven surfaces), and integration with the vehicle's drive-inhibit system.
Patent filings by OEMs are a noisy signal — BYD, like all major manufacturers, files aggressively and not every patent becomes a feature. However, BYD's track record of rapid feature-to-production cycles and its vertically integrated supply chain reduce the typical lag. The regulatory tailwind also exists: NHTSA and Euro NCAP have both increased scrutiny of low-speed pedestrian and vulnerable-road-user scenarios.
Open questions: sensor suite (camera-only vs. thermal vs. ultrasonic fusion), false-positive rate in real-world parking environments, and whether the system requires specific hardware or can be retrofitted via OTA to existing platforms with under-vehicle cameras.
Reality meter
Why this score?
Trust Layer BYD has patented an AI-based baseline imaging system capable of detecting living beings beneath a parked vehicle before it moves.
BYD has patented an AI-based baseline imaging system capable of detecting living beings beneath a parked vehicle before it moves.
- BYD has filed a patent on the technology, indicating a formal intellectual-property claim rather than a concept announcement.
- The system uses a baseline imaging approach — capturing a reference state of the undercarriage to compare against at departure.
- The stated detection target is 'living beings,' implying the system is designed to prevent low-speed under-vehicle collisions with people or animals.
- The source is a patent disclosure, not a product launch — no confirmed ship date, vehicle model, or production timeline is provided.
- The excerpt is thin on technical detail: sensor modality, false-positive rate, and real-world performance data are absent.
- Patent filings do not guarantee productionization; OEMs routinely file patents that never reach consumers.
A filed patent is a verifiable legal artifact, but the excerpt provides minimal technical or performance evidence to validate the system's real-world efficacy.
The signal type is correctly flagged as incremental — the source makes a specific, bounded claim about a patent, not a breakthrough, keeping hype moderate.
The problem being solved is real and under-addressed by current safety systems, but impact is contingent on production deployment, which remains unconfirmed.
- 1 source on file
- Avg trust 40/100
- Trust 40/100
Time horizon
Community read
Glossary
- baseline-differencing approach
- A detection method that captures a reference image of an area when it is confirmed clear, then compares subsequent images to identify changes or anomalies. In this system, it detects objects that appear under the vehicle between the initial stationary state and departure time.
- under-vehicle blind zone
- The area directly beneath and immediately around a vehicle that cannot be seen by the driver or standard safety sensors, creating a hazard during low-speed movement when pedestrians or objects may be hidden from view.
- liveness detection
- A technique that distinguishes living beings from inanimate objects by analyzing thermal signatures, motion patterns, or other biological indicators, rather than relying on shape or visual appearance alone.
- drive-inhibit system
- A vehicle safety mechanism that prevents or suppresses engine/motor operation or movement when certain unsafe conditions are detected, such as the presence of an obstacle or pedestrian.
- OTA (Over-The-Air)
- A wireless software update method that allows vehicle features and systems to be updated or installed remotely without requiring the vehicle to visit a service center.
- surround-view cameras
- A multi-camera system that provides a 360-degree bird's-eye view of the vehicle's surroundings, helping drivers see blind spots during parking and low-speed maneuvers.
What's your read?
Your read shapes future topic weighting.
Your vote feeds topic weights, community direction and future prioritisation. Open community direction
Sources
Optional Submit a prediction Optional: add your prediction on the core question if you like.
Prediction
Will BYD ship a production vehicle with this under-vehicle living-being detection system enabled by end of 2026?