SOFITEC is an international aerostructures manufacturer, specialised in the development of integrated solutions for composite, metallic and assembly projects.

current obstacles

The carbon fibre parts involved in the SOFITEC use case are not particularly heavy. Nevertheless, their large sizes and the distance to be covered between working areas restrict the manipulation by only a single operator. Most of the time, two operators are needed for this procedure.

In this current scenario, only one operator is bringing added value performing the various actions needed; the second operator is there to help the other operator to hold the heavy parts.

The challenge is to replace the intuitive and evident cooperation between the operators by an active Human-Robot collaboration, in order to correctly coordinate the task, including force, vision, voice modalities.

Mobile Manipulator

SHERLOCK solutions

Specifically, SHERLOCK introduces a mobile dual-arm robot coordinating with a human operator in carrying around the workplace the large composite parts. The collaborative robot moves along the workshop in tandem with the operator and actively cooperates with him to accurately position the different pieces on the destination locations. This collaboration is allowed thanks to multiple capabilities added on the robot, such as a physical using dual-arm impedance control as well as indirect interaction via human guidance perceived through vision and voice.

As the robot cannot just follow the operator, it has been enhanced by 3D vision tools to independently navigate inside the shop floor. Besides, the robot interactively learns a funnel of trajectories to perform its own manipulation and installation, in synchronisation with the operator. Due to the high number of parts and process variability, machine-learning is a key technology that a lot of SHERLOCK developments are based up to.

Mobile Manipulator


  • safety

    guarantee of responsive interaction between the operator and the autonomous robot

  • meticulous collaboration

    accurate environment perception through 3D cameras for the robot to optimize production quality and lower error rate

  • machine-learning

    capacity of robots to perceive their workplace environment, adjust their behaviour and provide on-the-spot instructions