Understand the home
LiDAR, IMU, odometry, ToF, microphone arrays, and vision provide environmental and interaction signals.
TECHNOLOGY · FOR BUILDERS
An AI brain
built for the home.
We develop an emotion-aware embodied-intelligence brain that coordinates perception, mobility, BSLA companion decisions, family memory, and applications. Kangkang is the first product.
View architecture
The home is the hard part
A family robot must do more than talk.
Long-term use requires navigation through physical space, appropriate social boundaries, and trust.
Space and motionPrivacySafetyRelationshipsEmotion and companionshipLearning over time
System architecture
A four-layer technology stack.
Move through the home
SLAM, path planning, dynamic obstacle avoidance, differential drive, and auto-charging work together.
Decide how to accompany
BSLA turns behavior, emotional cues, language, and authorized memory into companion actions.
Complete the interaction
Proactive conversation, reminders, family participation, and OTA updates connect robot and app.
In-house embodied mobility
A 25 cm chassis for complex homes.
Mapping, localization, path planning, obstacle avoidance, and auto-charging support movement in suitable indoor spaces.
Multi-sensor fusion
Radar, IMU, and odometry support mapping, localization, and state estimation.
Dynamic path planning
Global and local planning adjust routes around everyday obstacles.
Clear limits
Stairs, high thresholds, and wet or slippery floors are not suitable.
BSLA companion decisions
Beyond conversation,
timing and restraint matter.
BSLA is not a single chat model. Rules and models work together with behavior, emotional cues, language, and authorized family memory before the robot acts.
Whether to respondDo you need me now?
Balance a desire to interact with the need not to interrupt.
When to respondIs this the right moment?
Consider agreed schedules and current context.
How to respondWhat should I say or do?
Coordinate language, expression, and movement.
BSLA supports everyday companion interaction and continues to improve through real use. It does not provide health assessment, professional care, or emergency response.
Multi-user family memory
One home. Multiple people. Three memory layers.
FAMILY MULTI-USER MEMORY
- 01Context memoryKeep the present conversation connected
- 02Intent memoryConsider what the person wants to express
- 03Long-term memoryConnect experiences within authorized settings
Understanding across people and time.
The family memory system serves different household members through context, intent, and long-term memory. Together, the three layers provide cues for companion decisions across ongoing interaction.
Data processing and deletion remain governed by current product settings and the official privacy policy.
Emotional companion brain · Research directions
From understanding to action.
Next-stage research connects a deeper understanding of people and relationships with action in the physical home.
Family companion decision engine
A brain that understands the household.
Building on three-layer, multi-user family memory, further research will connect identity, emotional cues, and relationship graphs to inform when and how to respond.
Lingdong mobility system
A brain and body that work together.
Further development will address low-light charging, map updates after layout changes, and navigation and recovery in narrow spaces, connecting decisions with adaptive action.
R&D roadmap
Enter the home, train the brain, then expand capability.
Robots in real homes
Refine mobility, conversation, and proactive companionship through real use.
A growing companion brain
Improve decisions, memory, family participation, and safety boundaries.
Emotion-aware computing
Explore brain authorization and an EPU emotion-computing chip.
Technology and product partnership