Warehouses in fashion and retail handle huge volumes, thousands of references and seasonal collections, all under growing pressure from omnichannel distribution. Many tasks, from unloading and sorting to processing e-commerce returns, are repetitive and physically demanding. They are also hard to automate: conventional systems work best with rigid flows and regular objects, while garments on hangers are a challenge of their own.
What is ARAL
ARAL (AI-driven Robotic Assistant for Logistics) is a project led by PAL Robotics to develop a humanoid robot that can work in warehouses designed for people, without changes to shelving or existing infrastructure. It starts from KANGAROO, our bipedal humanoid research platform, and aims to take it from TRL 4 to TRL 5 in its first phase.
How it will work
The project focuses on three areas:
- Perception and cognition: combining vision with touch and torque sensing to handle soft textiles, including tasks where the camera cannot see, such as reaching into a deep box. Vision-language-action (VLA) models will let the robot understand simple spoken instructions.
- Adaptive learning: training in simulation so the robot walks and manipulates hangers with garments in a stable way, and learning new tasks through VLA models and reinforcement learning (RL).
- Mechatronics and safety: actuators that sense and regulate force so the robot can work safely alongside people.
Collaborating with our retail partners
ARAL is being developed with two leading retail partners, both of which we already work with. Our StockBot robots already support automatic inventory in large-format retail stores, and we have run several proofs of concept together. Now we are taking a further step: both partners will help us prioritise use cases, define performance indicators and acceptance conditions, and give our engineers access to their distribution centres to study real operations.
What comes next
Until May 2027 we will work with our partners to define the use cases and indicators, build a test bed in our Barcelona lab, develop the hardware and software, compare the results with manual performance and run a controlled pilot in a real distribution centre.
This is only the beginning: we want ARAL to open the way for humanoid robots that work side by side with people and bring physical AI to more tasks and more logistics environments. We will share our progress here.
ARAL is funded by the European Union through the European Innovation Council (EIC).
