Force/torque sensors provide real-time measurements of the contact forces and torques at the robot’s end-effector. In assembly tasks, this lets the controller respond to how parts behave when they touch, guiding the motion with a compliant, spring-like behavior rather than rigid, hard contact. With this data, the robot can insert, clamp, or mate parts more safely and reliably by adjusting grip force, insertion force, and path in response to sensed interaction, and it can detect misalignment or unexpected resistance early to stop or correct the motion. This capability specifically addresses interaction with the environment, which is why it’s the best choice. In contrast, measuring only torque for tuning motor drivers ignores the broader interaction context; relying on visual feedback comes from cameras or vision systems, not force/torque sensing; logging temperature of joints is handled by thermal sensors, not force/torque sensors.

SCARA robots are built to move mainly in a single plane. They use a pair of rotary joints that sweep the arm in the horizontal plane, giving precise X and Y positioning with fast, stiff motion. Because the motion stays in that plane, SCARA is especially well-suited for insertion tasks and other planar assembly work. A jointed-arm robot, by contrast, stacks multiple rotary joints along the arm, which allows movement and orientation in three-dimensional space—reaching up and down, twisting, and maneuvering around obstacles. That extra freedom comes with more complexity and typically different speeds for 3D tasks. So the defining distinction is that SCARA is limited to planar motion with two axes, whereas a jointed-arm robot can operate in 3D with multiple axes.

Collision-free motion planning is typically tackled with strategies that either explore feasible regions of the robot’s configuration space or optimize a trajectory that avoids obstacles. Sampling-based planners like RRT and PRM handle high-dimensional spaces by randomly sampling configurations and building connections that stay away from collisions, which makes them very effective in cluttered environments and scalable to many joints. Optimization-based trajectory planning, meanwhile, casts the problem as minimizing a cost (such as time, energy, or smoothness) subject to constraints, including collision avoidance, and then computes a collision-free path or trajectory that satisfies those constraints. Together, these approaches are widely used because they provide practical ways to guarantee avoidance while accommodating complex robot models and environments. Brute-force search over all joint configurations becomes infeasible as the dimensionality grows, relying solely on reactive control lacks foresight for obstacles, and impedance control focuses on interaction forces rather than producing a complete collision-free plan.

Leadthrough programming shines because it lets a person on the shop floor teach the robot by physically guiding it through the task. The operator moves the robot along the path, and the controller records the positions and motions to create a teach path. No specialized computer programming knowledge is required, so maintenance staff or operators can learn quickly and set up or modify tasks without waiting for a programmer. This makes it easy to adapt to changes in parts or processes and reduces downtime waiting for code changes. The other options don’t reflect this advantage: regular production downtime isn’t a benefit of leadthrough, and it doesn’t rely on complex scripts written by programmers. Also, it doesn’t inherently guarantee the fastest cycle times; speed is highly task-dependent and often requires optimization beyond simple taught moves.