Humanoid Simulation
We build virtual environments using tools such as NVIDIA Isaac Sim to validate motions and tasks for humanoids, primarily Unitree platforms.
Using tools and models such as Isaac Lab and Isaac GR00T, we support motion and function development, AI model training and tuning, and transfer to physical hardware, tailored to your goals.
3 Development Areas
Simulation Environment
Recreate the geometry, joint configuration, and workspace of humanoids such as Unitree platforms using simulation environments such as NVIDIA Isaac Sim.
Configure floor, contact, and sensor conditions for repeatable tests of posture, locomotion, and object manipulation.
Motion & AI
Develop motions and functions for tasks such as maintaining posture, walking, and using arms and hands.
Refine controls and AI models for the robot and task, checking balance, contact, and task completion to guide improvement.
Sim to Real
Transfer validated motions and AI models to hardware according to the robot's specifications and available development interfaces.
Check differences in sensor and joint responses and refine behavior while gradually expanding the tested operating range.
Work Type
Review the target robot, desired motions, available data, and development environment.
Define tasks and validation scope, then choose evaluation methods such as posture stability and task success rate.
Import and refine the robot model in a simulator such as NVIDIA Isaac Sim, configuring joint ranges, physical properties, and sensors.
Build the workspace and task objects into a virtual environment suited to the intended tasks.
Develop posture control, locomotion, and object manipulation for the robot's specifications and goals.
Test tasks combining multiple motions and identify loss of balance, contact, and failure conditions.
Choose methods suited to the robot and task, such as simulation-based learning with Isaac Lab or adapting Isaac GR00T foundation models.
Use collected motion data and simulation results to develop and tune AI models.
Compare behavior under consistent conditions and refine models based on success rate, stability, and execution speed.
Prepare hardware interfaces, execution conditions, and test procedures to transfer validated motions and models.
Check differences from simulation and feed hardware results back into the environment and model refinement.
EXDREAM R&D / Robot Decision System
Robot Decision Data
Our Robot Decision System (RDS) creates decision-learning data from humanoid motion logs, operator histories, and simulation results.
We are developing company- and task-specific decision datasets that capture states, candidate actions, selected actions, and outcomes.
Robot Decision Model
We are researching and developing lightweight AI models that evaluate and select the next action using the prepared decision data.
Through training and tuning for company-specific procedures and goals, we aim to support humanoid task decisions.
Robot Decision Harness
We are researching and developing mechanisms that adapt decision methods and compute to task complexity, risk, and the current state.
We compare combinations of models and decision methods in simulation, evaluating success rates, decision latency, and failure conditions.
DELIVERABLES
Deliver the simulation environment, including the robot model, workspace, and task configurations, in an agreed format.
Include execution conditions and operating and configuration guidance for continued validation.
Deliver developed and refined motions, control software, and AI models with validation conditions and results.
Define the included functions, motion logs, and training and evaluation data according to the project scope.
Organize hardware execution settings, interfaces, validated behaviors, and remaining checks.
Hand over the development environment and operating procedures so your team can continue testing and improving, with further refinement support where needed.
Step
Review the robot, intended tasks, and available data to define the development scope and evaluation criteria.
Prepare the robot model and workspace and configure joints, physical conditions, and sensors.
Develop the motions, functions, and AI models required for the task and run them in simulation.
Vary test conditions to check stability and task completion, then refine motions and models.
Test incrementally on hardware, adjust for differences from simulation, and hand over for continued development.
3 OFFERING TYPES
Based on your team structure and your project’s scope, scale, and goals, you can choose from the following three engagement models.
Around once a week
1–3 people
We create development plans for each company’s projects, document requirements, and support development team formation and vendor negotiations.
We also review actual progress and provide support to keep development projects running smoothly.
Around three days a week
1–3 people
We join your development team as your tech lead and work alongside you on design and development.
We share practical expertise and provide educational content, including seminars, so your employees can continue development independently after the project ends.
Embedded on site
Up to 5 people, depending on project scale
EXDREAM’s development team works on site at your company, handling everything from planning and coding to deployment.
For companies without engineers, we serve as your dedicated development team.
For companies with an existing engineering team, we join forces and work together toward successful delivery.
Services, team size, engagement frequency, duration, and responsibilities can be flexibly tailored to each project.
START A PROJECT