Robotics projects for engineering students: what to run on a real cobot

• Pick and place. Move parts between two positions, then between several, and measure how repeatable the motion is. It teaches coordinate frames, waypoints and speed settings.
• Gripper selection. Give students three different parts and ask them to choose and test a gripper for each. The lesson is that the tool decides what the robot can do.
• A small assembly line. Chain a cobot with a conveyor and a sensor so that a part arrives, is picked, and is placed in a fixture.
• PLC handshake. Let a PLC tell the cobot when to start and receive a signal when it finishes. Students see how a robot sits inside a larger control system.

Drag-to-teach is a good first session for all of these, since students can record a path by moving the arm by hand before they write any code.

Intermediate projects with vision and force

Once students can move the arm reliably, the Zeta42 page points to cobot programming, vision and force control as the core of engineering and computer-science degrees. Projects at this level include sorting parts by colour or shape with a camera, correcting a pick position when the part is not where it should be, and pressing or inserting a part with a limit on how hard the arm may push.

These projects suit computer-science students as much as mechanical and electrical ones, because the hard part is the software that links what the camera sees to where the arm goes. Mixed teams of the two groups tend to do well here.

Final year and research projects

For final year work, access matters more than polish. The Zeta42 page puts it this way: a platform that hides its control stack is a demonstration rather than an instrument, and a project that cannot reach below the interface is a project about the interface. An open Python API and ROS support let a student change how the robot is controlled rather than only use it.

Project directions the page lists include research on humanoids and data-collection rigs. A student might collect demonstrations from a bimanual cell, build a dataset from them, and train a model that repeats the task. A teaching server on the local network can run the models for a whole bench, so students see the weights and the reasons it slows down rather than calling a remote service.

Train faculty before you hand out projects

A department that has never run a cobot cannot supervise a project on one. The Zeta42 sequence is staff, then curriculum, then students, with operator and drag-to-teach workshops for the people who will supervise and SDK onboarding for the research group. Workshops run at the Abu Dhabi campus or on yours. For students on the operator track there is the DOBOT Certificate, and the STEM and robotics training page describes the courses.

How a department can plan it

Zeta42 describes a campus project in five steps, faculty first: map the courses, cohorts and research groups; design the bench layout, safety zones and platform mix; supply and install; train faculty; then certify and support. Zeta42 is DOBOT's partner for the UAE and the GCC, so the robots, PLC platforms, grippers, vision kits and rigs come from one supplier, with service and spares from Abu Dhabi. The page also notes that 405 universities and institutes worldwide are named in published research using DOBOT robots, which is a figure you can check against the papers themselves.

If you are planning a lab around these projects, see our robotics for universities and colleges page, then tell us whether teaching or research is the more urgent problem and we will start there.