US-based robotics company Figure announced that it has made an important progress in the production process of its Figure 03 humanoid robots, moving from the prototype stage to mass production. This transformation carried out at the BotQ production facility in California, increased the daily production capacity to one robot per hour.

This 24-fold increase in less than four months was made possible by more than 150 networked workstations and custom production software. The company stated that it successfully completed the delivery of more than 350 units during this period.

Production capacity and quality control processes

Figure optimized the production cycle by installing dedicated production lines for all critical modules of the Figure 03 robot at the BotQ facility. Strict inspections and more than 50 intermediate control points implemented to improve supplier quality directly affected production efficiency.

As a result of these improvements, the first pass efficiency at the end of the production line increased to over 80 percent. An efficiency rate of 99.3 percent was achieved in battery production and more than 9,000 actuators have been produced to date.

Each robot undergoes more than 80 functional checks, including stress tests such as squats and runs, to rule out early-stage failures. The company states that its expanding robot fleet accelerates the large-scale data collection process for the humanoid artificial intelligence model called Helix.

This growth in operational scale accelerates fault analysis, while software-based backup systems ensure that robots continue to operate in the event of minor errors. Figure aims to constantly monitor and keep all robots up to date by establishing in-field service, fleet management and wireless update systems.

Advanced sensing and movement capabilities

Figure also made an important update to its Helix System 0 (S0) artificial intelligence model, which increases the mobility of its robots.

Robots, which previously only understood joint states and body movements, can now perceive the environment in 3D using visual data from stereo cameras.

This new system allows robots to move more stably in complex environments such as stairs or irregular surfaces. These behaviors trained in a simulation environment can be directly transferred to the real world without the need for any additional calibration.

Robots can now walk on different surfaces with human-like stability, even in changing light conditions. This technological advancement is seen as a critical step towards Figure’s goal of developing fully autonomous and perception-driven control systems.

What do you think will change the mass production of humanoid robots in our daily lives?