Parallel Microrobot localised in a large workspace – micro-SPIDER

The partners have adopted a project management methodology following a V-cycle which makes it possible both to design and validate the overall system but also each sub-assembly. At the start of the project, the task T1 "Systemic approach of the microrobot" has the initial objective of providing specifications from this system approach to tasks T2, T3 and T4, respectively in charge of designing, modeling and validating the subassemblies "Actuation system", "parallel structure" and "measuring systems". In a second phase of the project, the task T1 will aim to validate the integration and characterize the performance achieved by combining the developments made in tasks T2, T3 and T4.

This structuration, combining a global vision of the system developed with a division into sub-assembly development tasks, makes it possible both to facilitate interactions between the different tasks and also to manage the potential risks existing in each sub-task by proposing adapted fallback solutions that do not block the overall progress of the project.

The work program of the µSPIDER project is structured in 5 tasks and will be spread over two phases: a first phase of establishing the specifications, then of design and sizing of the various subsystems (actuation surface, parallel structure, and measurement system) and a second phase of control, integration and characterization of the various experimental devices.

The results of the µ-SPIDER project will be of a theoretical and experimental nature.

• Multiphysics model at cell level then a global model of the surface

• Experimental prototype of the smart surface designed and sized from the model, taking into account the constraints imposed by the microrobot.

• Physical model of the microrobot

• Experimental 6-DoF microrobot prototype using commercially available short stroke actuators

• 3-DoF microrobot capable of bending over 90° in an extremely small space

• Modeling and sizing method of a network of Hall-effect sensors

• Experimental prototype of a smart surface instrumented by a network of Hall effect sensors

• Modeling and design of an optical tracking system for the localization of several pallets

• Experimental prototype showing the performance of the measurement by optical tracking

• Measurement methodology exploiting the fusion of optical and magnetic measurements

• Vision-based force estimation algorithms for continuum parallel robots

Several challenges were addressed during the micro-SPIDER project. They provide ideas for future work.

The integration of the displacement smart surface with the miniature parallel structure could not be completed due to time constraints and incompatibility between the dimensions of the parallel structure and existing actuators. An alternative actuation system is currently being studied as part of a doctoral thesis at the Roberval laboratory.

Furthermore, the 3D assembly of small-scale continuously deformable structures remains a major technological challenge. Methodologically, the calibration of these robots—essential to ensuring their accuracy—also poses difficulties, as there are many parameters that are challenging to measure at small scales. The measurement systems developed in the micro-SPIDER project (for position and force), combined with the use of finite element models, have helped advance the identification of geometric parameters for continuum parallel micro-robots.

The Hall-effect sensor network, designed and experimentally tested, has shown promising potential and will continue to be developed. However, it has also demonstrated that the calibration process is time-consuming in order to ensure high measurement accuracy over a large surface area. This has been identified as a challenge that will be the subject of further research at the Roberval laboratory.

Both partners are part of a consortium that has submitted a thematic project on miniature robotics within the PEPR Robotics framework. The work carried out in the micro-SPIDER project serves as a foundation for certain proposed initiatives, particularly in flexible robotics and small-scale measurement.

An ANR project (Xtremloc) was submitted and selected in 2024. The goal of this project is to develop a high-resolution spatial localization system dedicated to extremity surgery, particularly for the hand. This localization system is based on the technology developed in the micro-SPIDER project. It will be miniaturized and enable 3D localization.

Recent developments in microrobotics have focused on increasing the accuracy and dexterity of micromanipulators. Nevertheless, these robots remain very bulky compared to the size of the handled objects that induces a very unfavorable transported mass/moving mass ratio. The mass of these robots does not allow them to reach industrial rates either. A major challenge therefore concerns the reduction of the moving mass of micromanipulators. In addition, their workspace is often reduced because, on one hand, the precision of required performances often leads to a limitation of the actuator strokes and, on the other hand, the required measurement resolutions are generally obtained by sensors with small measurement range.
The objective of the micro-SPIDER project is to develop a new generation of micromanipulator combining a great dexterity (6 degrees of freedom), very low masses in movement and a micrometric precision over very large strokes. For this purpose, we propose to combine the know-how of the ROBERVAL laboratory in the fields of actuation and non-contact measurement with those of the FEMTO-ST institute in parallel microrobotics. This new micromanipulator generation is based on the use of a parallel structure whose feet will be mounted on mobile pallets. These pallets will be placed on an electromagnetic smart surface to move them in two (XY) or three degrees (XY?) of freedom over very large strokes. Since the movements of the pallets are limited to the plane, they are not able to perform directly three-dimensional micromanipulation tasks. To increase dexterity and in particular to obtain out-of-plane rotation movements, we propose to use and combine the movements of three pallets to actuate the parallel structure. An adapted deformable parallel architecture would indeed make possible to generate the six degrees of freedom of space from planar translation movements. As the actuation is performed by the smart surface, the parallel structure will be very light. The approach proposed here is innovative compared to current approaches that use more massive moving structures. Finally, parallel robots are known to have relatively small workspaces however, associated with a long-stroke smart surface, a parallel structure could reach a very large operational volume. This large range feature requires the development of a dedicated position measuring system of the parallel structure regardless of its position on the smart surface. The partners of the micro-SPIDER project propose two measurement approaches having in common the ability to perform a high-resolution measurement over a large planar range: a network of Hall effect sensors integrated into the smart surface, taking advantage of the magnetic property of pallets and a non-contact optical localization and tracking system with little impact on moving parts. Both measurement means will be compared in terms of their respective performances and measurements from these two systems can also be merged to improve the overall performances of the micromanipulator and streamline the deployment of these sensors.
The design and development methodology proposed in the micro-SPIDER project is also original since it couples a system approach allowing optimizing the overall design and performances of the proposed micromanipulator as well as a disciplinary approach (parallel structure, actuation system and measuring system) allowing to developing precisely each sub-system.