Automation Engineering

Automation engineering brings together mechanics, electronics, control theory and software to design, build and integrate systems that perform tasks with minimal human intervention. Its scope ranges from precision calibration of industrial robots and machine tools to autonomous path planning for aerial and ground vehicles, as well as the optimisation of large-scale manipulators driven by cables. Modern work in the field emphasises digital models and virtual commissioning, real-time networking of embedded controllers, and the use of data analytics and learning algorithms to enhance adaptability. Automation engineers tackle problems of geometric accuracy, dynamic performance, fault tolerance and multi-agent coordination, creating solutions that power smart factories, driverless transport, service robotics and beyond. By uniting engineered hardware with software architectures and communication standards, automation engineering underpins productivity gains, quality assurance and novel applications across manufacturing, logistics, environmental monitoring and infrastructure.

Research from Nature Portfolio

A novel compensation scheme for geometric errors in grinding robots combines a Levenberg–Marquardt optimiser with a logistic–tent chaotic mapping step to explore a dense set of candidate solutions around initial estimates. This hybrid strategy minimises truncation error and diversifies the search, yielding a 6 % increase in accuracy and nearly 48 % speed-up in convergence, and demonstrating significant improvements in workpiece surface quality.

Researchers have introduced a kinematic calibration method for collaborative redundant manipulators by closing the kinematic loop with a spherical joint. Torque sensors provide constrained motion data, and a kinesthetic approach yields optimal parameter estimates via nonlinear optimisation. Trials on a seven-degree-of-freedom robot show marked gains in absolute positioning, matching the performance of optical systems at a fraction of the cost.

An enhanced local planning algorithm for multirotor flight exploits B-spline parameterisation and distance gradient information to smooth collision-free paths in unknown environments. A limited-memory BFGS solver iteratively refines an initial trajectory generated by obstacle avoidance, while a time-scaling step enforces dynamic feasibility. Simulation results reveal faster computation, shorter paths and reduced control effort compared with conventional spline-based planners.

Research from all publishers

Optimisation frameworks for reconfigurable cable-driven parallel manipulators determine cable-attachment locations that maximise wrench-closure and tension margins throughout a specified workspace. By formulating wrench-feasible conditions as inequality constraints over attachment coordinates, semidefinite programming and gradient-based solvers compute configurations that balance force distribution and minimise actuator loads in both low- and high-degree-of-freedom platforms.

PANTHER, a perception-aware trajectory planner, jointly optimises UAV rotation and translation to keep dynamic obstacles within the sensor field of view and minimise image blur for robust tracking. By exploiting the differential flatness of multirotors and embedding under-actuated dynamics via a fibration mapping, the planner achieves higher success rates and obstacle observability in real-time hardware tests compared with decoupled yaw-translation schemes.

Automation Engineering publication trend

The graph below shows the total number of articles in automation engineering across all publications each year (not limited to Nature Index journals).

Technical terms

Levenberg–Marquardt algorithm: A nonlinear least-squares optimiser that interpolates between gradient descent and the Gauss–Newton method to refine parameter estimates in calibration problems.

Logistic–tent chaotic mapping: A method for generating diverse candidate solutions by iterating a hybrid of logistic and tent maps over an interval, enhancing global search capabilities.

Closed kinematic chain: An arrangement in which a manipulator’s end-effector is connected back to its base by one or more additional links or joints, improving stiffness and measurement accuracy during calibration.

B-spline trajectory: A smooth, piecewise-polynomial curve defined by a small number of control points, commonly used for path planning that meets continuity and dynamic constraints.

Cable-driven parallel manipulator (CDPM): A parallel robot in which the end-effector is actuated solely by multiple cables under tension, offering large workspaces and high payload-to-weight ratios.

Perception-aware planning: A trajectory optimisation approach that incorporates sensor field-of-view and measurement quality metrics into the cost function to improve real-world obstacle detection and tracking.

References

1. A logistic-tent chaotic mapping Levenberg Marquardt algorithm for improving positioning accuracy of grinding robot. Scientific Reports (2024).
2. Kinematic model calibration of a collaborative redundant robot using a closed kinematic chain. Scientific Reports (2023).
3. Gradient-based autonomous obstacle avoidance trajectory planning for B-spline UAVs. Scientific Reports (2024).
4. Cable Attachment Optimization for Reconfigurable Cable-Driven Parallel Robots Based on Various Workspace Conditions. IEEE Transactions on Robotics (2023).
5. PANTHER: Perception-Aware Trajectory Planner in Dynamic Environments. IEEE Access (2022).