Spent six months designing a single, non-modular cobot-style arm from scratch. Realized that any design change required a full rebuild, and after speaking with professors, saw the potential for a joint-based, modular approach instead. One joint library that could become hundreds of different robots. This insight became the origin of CORA.

CoDI (Cora Desktop Interface) had a first working version released. This started in the six months before the project but reached a significant functional milestone. Now users without Ubuntu or ROS can control the arm.

Figured out how the platform would work at both a software and hardware level: a library of distinct joint types, each built to a standardized interface, that could be combined together in different configurations to form a working manipulator. This concept became the foundation the rest of the platform was built around.

Recruited a team of talented individuals to take on the challenge of developing this open source project. The team at this point consists of two mechanical engineers and one mechatronics/software engineer The first phase involves continuing to work on and finishing the first prototype which had six months of work already put into it.

Designed a wolfrom style cycloidal drive as the first candidate actuator for the joint library. The design was ordered and built, but the actuator did not perform as intended, so the approach was later set aside in favor of alternative drive designs.

Allied Motion, a motor manufacturer, reached out to request a demo model showcasing their motors integrated into CORA’s joints. The connection came through a representative from Breda Robotics, who attended one of RSA’s project evenings and introduced the opportunity to the team.

First meeting with allied motion to discuss their vision and requirements. we had overall made a very good impression and they offered us a full sponsorship for the collaboration. It is now up to us to make an appropriate motor selection based on the task and relay this information back to them

Established CAN bus communication between joints and got the ROS 2 Jazzy control stack operational with real encoder feedback on a working prototype joint, validating that the standardized joint interface works reliably in practice.

Development was started on a new tool that will empower users to design their own arm in a 3D viewer, then export required modular packages and documents in order to be able to build and control their arm. This also enabled us to think deeper about the software architecture and what exactly needs to be static and what doesn’t

Finalized the design of the back shells a week prior. For our first 6-dof prototype, six joints were then promptly soldered correctly and assembled.

The first prototype has been assembled and is ready for show during the Q1 RSA project evening. After solving many issues with communication and calibration the joints finally had the will to be assembled and put to use.

Complete the CORA Configurator, automatically generating the CAD assembly, ROS 2 software stack, and simulation environment from a user’s configuration file, so future arms can be built without manual integration work.

Demonstrate the full pipeline on a large scale live at [Undisclosed event]. The user configures an arm, the software generates everything needed to build and run it, and the resulting robot is assembled and operated on the spot.

Publicly release the stable version CORA, free of hardware and software bugs including a full documentation site and build guides, aswell as working configurator and CoDI. Expand the joint library so external student teams and researchers can build their own manipulators on top of the platform rather than starting from scratch. This also kickstarts a massive push in terms of outreach. At this point awareness of our project in european robotics and beyond is necessary to really have a positive impact.