Articulated robot
An articulated robot is a type of industrial robot defined as an automatically controlled, reprogrammable manipulator with at least three rotary joints in its arm, enabling a wide range of motion and flexibility similar to a human arm.[1][2] These robots are classified by the number of axes or points of rotation, with the most common configuration being six axes, which allows for precise positioning and orientation in three-dimensional space.[3] Unlike other robot types such as Cartesian or cylindrical robots, articulated designs use serial revolute joints without kinematic constraints, providing superior dexterity for complex tasks but requiring advanced control systems to manage their nonlinear kinematics.[4] The origins of articulated robots trace back to the mid-20th century amid the rise of industrial automation, with early developments building on the first industrial robot, Unimate, introduced in 1961 for materials handling at General Motors.[3] A pivotal advancement came in 1969 with the Stanford Arm, the first all-electric, computer-controlled six-axis articulated robot, which laid the foundation for modern electric servo-driven systems and enabled programmable precision in manufacturing.[5] Since then, standards like ISO 8373 have formalized their classification, emphasizing reprogrammability in three or more axes for multipurpose industrial use, driving widespread adoption through the 1970s and beyond as computing power improved robot control and safety.[1] Articulated robots excel in versatility, with sealed joints and protective features allowing operation in harsh environments, and their mountability on floors, ceilings, or rails expanding deployment options in factories.[3] Key applications include arc welding, assembly, machine tending, painting, and material handling such as palletizing or loading, where their long reach and ability to navigate non-parallel planes outperform rigid-coordinate robots.[2][1] While they offer high precision along complex trajectories, their higher cost and mass limit use in ultra-high-speed scenarios compared to simpler designs.[3]Definition and Characteristics
Definition
An articulated robot is a manipulator with three or more rotary joints, as defined by ISO 8373.[6] It is a type of industrial robot featuring multiple rotary joints, known as revolute joints, arranged in a serial kinematic chain to enable flexible, multi-dimensional movement. This structure allows the robot to mimic the dexterity and reach of a human arm, with rigid links connecting the joints to form a manipulator capable of navigating complex paths within its workspace. Typically equipped with 4 to 6 axes of rotation—though models with up to 10 axes exist—these robots are classified by their points of rotation, with the 6-axis configuration being the most prevalent in industrial applications.[3][7][8] In contrast to Cartesian robots, which rely on linear prismatic joints for precise but restricted straight-line motions along orthogonal axes, articulated robots use revolute joints to achieve superior reach and the ability to maneuver around obstacles or across non-parallel planes. Similarly, while SCARA (Selective Compliance Articulated Robot Arm) robots incorporate revolute joints for horizontal flexibility with selective compliance, they are limited to 4 axes and primarily operate in parallel planes, lacking the full three-dimensional versatility of standard articulated designs. This revolute joint emphasis provides articulated robots with enhanced adaptability for tasks requiring varied orientations, though at the expense of higher complexity and cost compared to these alternatives.[3][3] The anthropomorphic nature of articulated robots stems from their serial manipulator architecture, which parallels the human upper limb through sequential joints at the base (waist rotation), shoulder (pitch and roll), elbow (pitch), and wrist (yaw, pitch, and roll), connected by forearm and upper arm links. This configuration supports a broad workspace envelope and high degrees of freedom, typically 6 for full pose control in space. The Stanford Arm, developed in 1969, was the inaugural all-electric, 6-axis design that demonstrated programmable arm solutions for assembly tasks.[7][9][10]Key Features
Articulated robots feature a structure composed of segmented links connected by multiple rotary joints, enabling a high degree of flexibility and mimicking the human arm's configuration.[11] These robots typically support payload capacities ranging from 1 kg for small models to over 1000 kg for heavy-duty industrial variants, allowing them to handle diverse tasks from precision assembly to material transport.[12] Their reach generally spans 0.5 to 5 meters, determined by the arm's length and joint configuration, while repeatability— the ability to return to the same position—achieves precision levels of 0.01 to 0.1 mm, essential for applications requiring consistent accuracy.[12][13] The mobility of articulated robots derives from their rotational freedom across typically 4 to 6 axes, resulting in a spherical or hemispherical work envelope that encompasses a three-dimensional volume accessible by the end effector.[11] This design allows the robot to navigate complex paths and orientations within its workspace, providing versatility over linear or fixed configurations. Construction materials emphasize lightweight alloys such as aluminum for the arm segments to minimize inertia and enhance speed, paired with high-strength steel or composite components for the base to ensure stability under load.[14] Power sources for articulated robots predominantly consist of electric servo motors, which offer precise control and energy efficiency for most industrial models; however, hydraulic systems are employed in heavy-duty variants to deliver greater force and torque for payloads exceeding 500 kg.[15][16]History
Early Developments
Articulated robots emerged in the context of 1950s industrial automation efforts aimed at automating repetitive manufacturing tasks, with George Devol's 1954 patent for the Unimate serving as a key precursor. The Unimate, a hydraulic manipulator arm developed by Devol and commercialized through Unimation Inc., represented an early step toward programmable robotic systems but featured limited degrees of freedom, primarily using prismatic and revolute joints in a non-anthropomorphic configuration.[17] This design laid foundational concepts for industrial automation, though it was not a fully articulated, multi-axis system mimicking human arm motion.[18] A pivotal milestone in articulated robot development occurred in 1969 when Victor Scheinman, a mechanical engineering student at Stanford University, invented the Stanford Arm, the first electrically actuated six-degree-of-freedom (6-DoF) articulated robot. Unlike hydraulic predecessors, the Stanford Arm employed electric motors for precise control, enabling it to perform complex manipulations such as small-parts assembly through computer-based programming. This design introduced rotary joints at the shoulder, elbow, and wrist, allowing a wider range of motion and serving as a model for subsequent anthropomorphic robots.[19] Scheinman's innovation addressed limitations in earlier manipulators by providing closed-form kinematic solutions, marking a shift toward versatile, electrically driven articulated systems.[18] In the 1970s, Joseph Engelberger, co-founder of Unimation with Devol, advanced articulated designs by acquiring and adapting Scheinman's technology. In 1977, Unimation purchased the Stanford Arm design from Scheinman's Vicarm Inc., leading to the development of more refined articulated models like the PUMA series, which incorporated electric actuation for improved dexterity in assembly tasks. Engelberger's patents and engineering efforts during this period focused on enhancing control systems for these articulated forms, building on Unimate's hydraulic base to create commercially viable electric variants.[19] This transition was supported by early patents emphasizing joint coordination and programming, though implementation remained tied to Unimation's industrial focus.[17] Early articulated robots faced significant challenges, including limited positional precision due to mechanical backlash in joints and actuators, as well as high initial costs—which restricted adoption to large manufacturers like General Motors. These factors, combined with the need for specialized programming and maintenance, resulted in slow market penetration, with installations primarily limited to automotive die-casting and welding applications through the 1970s.[17] Despite these hurdles, such developments established the core principles of articulated robotics, paving the way for broader industrial integration.[18]Industrial Evolution
The commercialization of articulated robots accelerated in the late 1970s with the introduction of the PUMA (Programmable Universal Machine for Assembly) by Unimation in 1978, a six-axis electric manipulator designed for precise assembly operations. This model, developed in collaboration with General Motors, saw its first installations at GM facilities in 1979, marking a pivotal shift toward programmable, versatile robots suitable for diverse industrial tasks.[20][21] From the 1980s through the 2000s, articulated robots experienced robust growth, particularly through their integration with computer numerical control (CNC) systems, which enhanced automation in machining and material handling by allowing robots to perform loading, unloading, and coordinated operations with high accuracy. This period also witnessed market expansion beyond automotive manufacturing into electronics assembly, where articulated robots handled delicate component placement and soldering, driven by demands for miniaturization and speed in consumer electronics production.[22][23] Key milestones included the ascent of Japanese manufacturers, exemplified by FANUC's dominance in the 1980s; the company, having introduced its first all-electric industrial robot in 1974, captured over 20% of the global market by 1987, leveraging innovations in servo technology and reliability to lead in articulated arm production. Concurrently, the 1980s featured the formulation of international standards for industrial robots under ISO, including foundational definitions and safety guidelines that promoted interoperability and risk reduction across deployments.[24][25] In recent trends up to 2025, the maturation of collaborative robots (cobots) has transformed industrial applications, with Universal Robots launching its UR series in 2008—starting with the UR5 model—to enable safe, flexible operation alongside human workers without extensive safety barriers. This innovation contributed to surging adoption, as evidenced by global industrial robot installations surpassing 500,000 units annually by 2023 and reaching 542,000 units in 2024, reflecting broader technological integration and economic scalability.[26][27]Design and Mechanics
Joint Types and Configurations
Articulated robots predominantly feature revolute joints, which allow rotational motion around an axis, mimicking the flexibility of human limbs and enabling a wide range of orientations in three-dimensional space.[28] Prismatic joints, which permit linear translation, are rare in pure articulated designs, as they are more characteristic of Cartesian or cylindrical robots rather than the rotational emphasis of articulated structures.[4] The kinematic chain in articulated robots is typically a serial open-chain arrangement, consisting of interconnected links joined sequentially by these revolute joints, starting from a fixed base and progressing through shoulder, elbow, and wrist segments to the end effector. This serial structure provides versatility in reach and manipulation but can introduce challenges like limited workspace compared to parallel mechanisms.[29] Common configurations include the anthropomorphic setup, which emulates the human arm with six degrees of freedom (DoF) using an RR-RRR joint arrangement—two revolute joints at the shoulder, one at the elbow, and three at the wrist—for tasks requiring precise positioning and orientation.[30] The spherical wrist configuration enhances dexterity by incorporating three intersecting revolute axes at the wrist, allowing the end effector to orient in any direction without translation. Some articulated robot designs use offset wrist configurations, where the axes do not intersect, to improve manipulability and avoid singularities by preventing axis alignment that can restrict motion and lead to workspace limitations.[28] Redundant configurations, with seven or more DoF, extend the anthropomorphic design by adding an extra joint, often at the elbow or shoulder, to provide additional flexibility for obstacle avoidance and singularity evasion while maintaining the serial revolute chain.[32]Degrees of Freedom
In articulated robots, degrees of freedom (DoF) quantify the number of independent motions or parameters required to specify the configuration of the robot's end-effector, enabling precise control over its position and orientation in space.[33] For tasks in three-dimensional environments, a standard articulated robot arm typically possesses 6 DoF, comprising 3 translational degrees (along the x, y, and z axes) and 3 rotational degrees (about those axes), which allow full pose control of the end-effector relative to the base.[34] In serial chain configurations common to articulated robots, the total DoF is calculated as the number of joints n minus any constraints on their motion, though unconstrained serial chains with single-DoF joints per link generally yield DoF equal to n.[35] This formulation, derived from mobility criteria like Grübler's formula adapted for spatial mechanisms, underscores how joint count directly influences the robot's versatility.[35] The number of DoF significantly impacts performance, as higher values introduce kinematic redundancy, permitting alternative joint configurations to achieve the same end-effector pose while avoiding obstacles or optimizing trajectories.[36] For instance, the KUKA LBR iiwa features 7 DoF, leveraging this extra degree to enhance flexibility in collaborative settings by rerouting motion around hindrances without altering the task endpoint.[36] DoF also governs workspace analysis through manipulability measures, which assess the robot's ability to execute dexterous motions within its reachable volume; higher DoF expands manipulability by enlarging the ellipsoid representing feasible end-effector velocities for unit joint speeds.[37] These metrics, such as the condition number of the Jacobian matrix, highlight how DoF distribution affects ease of motion in different postures. A key limitation arises at singularities, configurations where the effective DoF temporarily reduces due to joint alignment, restricting instantaneous motion directions and potentially causing uncontrolled velocities or loss of task-space control.[38] For example, when an articulated arm is fully extended, the Jacobian matrix becomes rank-deficient, dropping the instantaneous DoF and complicating precise maneuvering near workspace boundaries.[38]Components
Actuators and Drives
Articulated robots primarily rely on electric servo motors as actuators to drive their revolute joints, providing precise control and high repeatability essential for tasks requiring accuracy. These motors, typically DC or AC types, convert electrical energy into mechanical torque and are the most common choice due to their compact size, low inertia, and ability to integrate with feedback systems for position control.[39][16] For applications demanding higher force, such as heavy material handling, hydraulic actuators are employed, utilizing pressurized fluid to generate substantial power output, though they are bulkier and require maintenance for seals and fluid systems.[40] Pneumatic actuators, powered by compressed air, are used in scenarios prioritizing speed over precision, offering rapid motion but limited control due to compressibility of air.[40] Drive systems in articulated robots often incorporate gear reducers to amplify torque from the actuators while reducing speed, with harmonic drives being particularly prevalent for their zero-backlash performance and high gear ratios up to 160:1, enabling smooth and accurate joint motion without play.[41] These strain-wave gears consist of a wave generator, flex spline, and circular spline, providing compact, lightweight solutions ideal for multi-axis arms.[42] Belt drives, using timing belts and pulleys, are alternatively selected for joints where higher speeds are needed and backlash tolerance is acceptable, offering quieter operation and easier maintenance compared to geared systems.[43] Key specifications for these actuators and drives include torque outputs reaching up to 6000 Nm in large industrial models for base joints handling heavy payloads, and maximum joint speeds of up to 500°/s for wrist rotations to support fast cycle times.[44][45] Energy efficiency is enhanced in modern electric servo systems through features like regenerative braking, which recovers kinetic energy during deceleration and feeds it back to the power supply, reducing overall consumption by up to 30% in repetitive tasks.[46] Selection of actuators and drives for articulated robots involves matching torque and speed capabilities to the required payload—typically 5-800 kg depending on arm size—and desired cycle times, ensuring the system can accelerate and decelerate without exceeding thermal limits or reducing lifespan.[47] For instance, precision assembly may prioritize low-backlash harmonic drives with efficient servo motors, while high-throughput welding favors pneumatic or belt-driven setups for quicker joint movements.[48]Sensors and End Effectors
End effectors are the terminal components of articulated robots that enable direct interaction with the environment, such as grasping, welding, or material handling. Common types include parallel-jaw grippers, which use two opposing fingers to securely hold objects through mechanical clamping; weld torches for precision arc, gas, or spot welding tasks; and suction cups that create vacuum adhesion for non-porous surfaces like glass or metal sheets.[49][50][51] To facilitate rapid tool exchange and adaptability in production lines, quick-change systems adhere to standards like ISO 9409-1, which specifies a circular plate interface with defined dimensions, mounting holes, and markings for mechanical compatibility between the robot flange and end effector. This standardization ensures exchangeability of hand-mounted tools, reducing setup time and enhancing modularity in articulated robot designs.[52] Sensors provide essential feedback for precise operation and safety in articulated robots. Joint encoders, typically rotary or absolute types, measure angular positions at each joint to enable accurate tracking of the robot's configuration and closed-loop control. Force and torque sensors detect applied pressures and rotational forces, allowing robots to achieve compliant behavior by adjusting motions to avoid damage during delicate tasks like assembly. Vision cameras, often 2D or 3D systems, facilitate object detection by processing images to identify shapes, positions, and orientations in real-time.[53][54][55] Integration of sensors enhances end effector performance, particularly through tactile sensors that enable slip detection by monitoring shear forces or vibrations at the contact point. For instance, Schunk grippers incorporate embedded force/torque sensing to measure and process interaction forces, supporting adaptive gripping that prevents object loss during manipulation.[56][57] Sensor fusion, combining data from multiple sources like encoders, force/torque, and vision, is critical for calibration and overall accuracy, especially in collaborative robots developed since the 2010s, where it improves manipulation precision through refined environmental perception and error compensation.[58]Kinematics and Control
Forward Kinematics
Forward kinematics is the process of determining the position and orientation (pose) of the end-effector in an articulated robot given the values of its joint variables, which represent the degrees of freedom of the mechanism. This mapping from joint space to Cartesian space is essential for understanding the robot's reachable workspace and is achieved through a chain of coordinate transformations that account for the geometry and joint configurations of the serial manipulator.[59] The standard approach for computing forward kinematics in serial articulated robots employs the Denavit-Hartenberg (DH) convention, which systematically assigns coordinate frames to each link and defines the spatial relationship between consecutive frames using four parameters: the link length a_i (distance along the common normal between joint axes z_{i-1} and z_i), the link twist \alpha_i (angle between z_{i-1} and z_i about the common normal x_i), the link offset d_i (distance along z_{i-1} from the origin of frame i-1 to the intersection with x_i), and the joint angle \theta_i (angle between x_{i-1} and x_i about z_{i-1}). These parameters enable the representation of the transformation from frame i-1 to frame i as a 4×4 homogeneous transformation matrix A_i, which combines rotation and translation in a single matrix form suitable for serial chains. The DH parameters are determined based on the robot's physical structure, with \theta_i or d_i varying as the joint variable for revolute or prismatic joints, respectively, while the others are fixed. This convention was originally proposed for analyzing lower-pair mechanisms, including robotic manipulators.[60][61] The general form of the A_i matrix is: A_i = \begin{bmatrix} \cos \theta_i & -\sin \theta_i \cos \alpha_i & \sin \theta_i \sin \alpha_i & a_i \cos \theta_i \\ \sin \theta_i & \cos \theta_i \cos \alpha_i & -\cos \theta_i \sin \alpha_i & a_i \sin \theta_i \\ 0 & \sin \alpha_i & \cos \alpha_i & d_i \\ 0 & 0 & 0 & 1 \end{bmatrix} The overall transformation matrix T from the base frame to the end-effector frame for an n-joint serial robot is obtained by multiplying the individual link transformations: T = A_1 A_2 \cdots A_n. The upper-left 3×3 submatrix of T represents the orientation (rotation matrix), while the rightmost column (excluding the bottom 1) gives the position vector of the end-effector origin.[59][61] For a typical 6-degree-of-freedom (6-DoF) articulated robot arm, such as those used in industrial manipulation, the DH parameters are assigned to each of the six links based on the specific geometry (e.g., anthropomorphic configuration with revolute joints). The forward kinematics derivation involves constructing the six A_i matrices—each incorporating the fixed parameters a_i and \alpha_i from the robot's design, and the variable \theta_i for each revolute joint—and computing their product T = A_1 A_2 A_3 A_4 A_5 A_6. This yields the end-effector pose in closed form, though the explicit position and orientation equations are nonlinear trigonometric functions of the joint angles, often left in matrix form for computational efficiency in control systems. The resulting T matrix fully describes the 6-DoF pose, enabling precise task planning without solving for each component separately.[59] As an illustrative example, consider a simple 3-DoF planar articulated arm with all revolute joints in the xy-plane, link lengths l_1 = 2 m, l_2 = 1.5 m, l_3 = 1 m (corresponding to a_1 = l_1, a_2 = l_2, a_3 = l_3), and all \alpha_i = 0^\circ, d_i = 0 m for simplicity. The DH table is:| Link i | a_i (m) | \alpha_i (°) | d_i (m) | \theta_i (°) |
|---|---|---|---|---|
| 1 | 2 | 0 | 0 | \theta_1 |
| 2 | 1.5 | 0 | 0 | \theta_2 |
| 3 | 1 | 0 | 0 | \theta_3 |