Robotic leg actuator, robotic leg comprising said actuator

Abstract

A robotic leg actuator (20), comprising: a speed reducer (30) having an input shaft (70) and an output element (46); the speed reducer (30) comprising at least one cycloidal disk (60) configured to be driven in an orbital motion by the input shaft (70); a plurality of output pins (84) connected to the output element (46) and passing through corresponding holes (86) in the at least one disk (60); and a plurality of output bushings (88), each bushing being disposed within one of said holes (86) to provide an interface between the at least one cycloidal disk (60) and one of said output pins (84), wherein each output bushing (88) is configured to provide a wear-resistant outer surface to cooperate with the cycloidal disk (60) and a low-friction inner surface to cooperate with the output pin (84).

Description

Robotic leg actuator, robotic leg comprising said actuator Technical field This document relates to a robotic leg actuator and a legged robot comprising said actuator. Previous technique Legged robots, like bipedal and quadruped robots, are increasingly used for complex tasks in challenging environments. The performance of these robots—their agility, speed, efficiency, and ability to interact with their surroundings—depends heavily on the design of their joint actuators. An actuator for a robotic leg must combine several demanding characteristics: high torque density to provide powerful movements, low weight and a compact form factor to minimize inertia and energy consumption, high efficiency, and a high degree of recoil capability to enable compatible control and efficient energy recovery. Cycloidal gears, or cycloidal reducers, are a well-known class of speed reduction mechanisms often used in robotics due to their ability to provide high torque and high reduction ratios in a compact package. Their operating principle, based on rolling contact rather than sliding friction, makes them robust and resistant to shock loads, which are common during locomotion. To achieve stable, low-vibration operation, conventional cycloidal gearboxes are typically designed with a double-disc configuration. In this arrangement, two identical cycloidal discs are mounted on a double-eccentric input shaft, offset by 180 degrees from each other. The orbital motion of the first disc generates an inertial force that is continuously counteracted by the equal and opposite inertial force of the second disc. This dynamic balance is effective but comes at a significant cost in terms of mechanical complexity. The dual-disc architecture inherently requires a greater number of parts, including two cycloidal discs, a more complex double-eccentric shaft, and at least two sets of bearings for the discs. This results in greater axial length, increased weight, and higher manufacturing and assembly costs. For applications where minimizing weight and size is critical, such as in the distal joints of a robotic leg, the complexity and mass of a conventional dual-disc cycloidal gearbox can become a significant drawback. Furthermore, integrating an actuator into a robotic joint often requires additional external bearings to support the limb's structural loads. The output bearing of a standard actuator may be sufficient to support its own output shaft, but not the significant radial, axial, and tipping moment loads experienced by a robot's hip or knee joints during dynamic movements. This need for separate joint bearings adds further complexity, weight, and bulk to the overall robot design. Therefore, there is a need for a robotic actuator that retains the high torque and robustness advantages of a cycloidal reducer while overcoming the limitations of conventional designs in terms of complexity, size, and weight. There is also a need for an actuator that simplifies its mechanical integration into a robotic limb, providing not only torque but also structural support for the joint itself. Summary The purpose of this document is to overcome the aforementioned drawbacks by providing a robotic leg actuator comprising a cycloidal reducer, which comprises: an eccentric input shaft; a cycloidal disk operatively coupled to the eccentric input shaft, such that the rotation of the input shaft imparts an orbital motion to the cycloidal disk, said orbital motion generating an unbalanced dynamic force; and an output element operatively coupled to the cycloidal disk; The actuator is characterized in that the cycloidal reducer further comprises a counterweight rotatably connected to the eccentric input shaft and configured to counteract the unbalanced dynamic force generated by the orbital motion of the cycloidal disc. The term "robotic leg actuator," as used herein, refers to a module configured to generate controlled rotary motion at a joint of a robotic leg. The actuator may comprise a motor, a speed reduction transmission, and control electronics, all housed within a single enclosure, and may be designed to provide precise torque and position control for movements such as flexion, extension, or abduction of a robotic limb. The term "cycloidal reducer" refers to a type of gear mechanism that converts the high-speed, low-torque rotation of an input shaft into the low-speed, high-torque rotation of an output element. Its operation may be based on an eccentric mechanism that drives one or more discs in a cycloidal fashion along an orbital path, causing them to mesh with fixed external pins or a toothed ring, thus producing a reduced-speed output rotation. The term "eccentric input shaft" denotes a shaft that has at least one journal or section, called an eccentric, whose central axis is parallel to, but offset from, the main axis of rotation of the shaft. When the shaft rotates, this eccentric section describes a circular path, thus imparting orbital motion to any component mounted on it, such as a bearing and a cycloidal disk. The term "cycloidal disc" refers to a plate-like component, for example, with an outer periphery profiled with a series of lobes, often in the shape of an epitrochoid. The disc may be configured to be driven in an orbital motion by the eccentric input shaft and to engage with a set of fixed pins or rollers to achieve speed reduction. It may also be provided with features, such as through holes, to transfer its resulting low-speed rotation to an output element. The term "operatively coupled," as used herein, denotes a functional connection between two or more components. This connection may be direct, for example, by direct physical contact or rigid attachment, or indirect, through one or more intermediate components, provided that it allows one component to influence the movement or function of the other, as described. For example, the cycloidal disk may be operatively coupled to the output element via output pins. The term "orbital motion" refers to the motion of a body, such as the cycloidal disk, in which its center of mass moves along a circular path around a central axis of rotation, without the body itself rotating at the same speed or in the same direction around its own center of mass. This motion is distinct from pure rotation, in which a body rotates around its own fixed center. The expression "the rotation of the input shaft imparts an orbital motion to the cycloidal disk" describes the kinematic relationship in which the rotation of the eccentric section of the input shaft forces the center of the cycloidal disk (which is mounted on said eccentric) to follow a circular path. The coupling of the disk's outer profile with fixed elements, such as annular pins, prevents the disk from rotating with the shaft, resulting in the desired orbital motion. The term "unbalanced dynamic force" refers to the time-varying inertial force, commonly known as centrifugal force, that arises when a mass is driven into orbital or rotational motion around an axis from which it is displaced. In the context of the actuator, this force is generated by the mass of the cycloidal disk and its associated components as they are driven into an orbital path by the eccentric input shaft. The term "output element" refers to the gearbox component that transmits the final, low-speed, high-torque rotary motion to the outside of the actuator. This element can take various forms, such as an output shaft, a flange with a mounting pattern, or a hub, and can be configured to connect to an adjacent robot limb or another mechanical component. The term "counterweight" refers to a mass intentionally added to a rotating or oscillating system to balance the forces generated by other moving masses. In this document, it may be a mass rigidly connected to the input shaft, positioned and sized to generate an inertial force that is substantially equal in magnitude and opposite in direction to the unbalanced dynamic force of the cycloidal disk. The expression "rotational connection to the eccentric input shaft" means that the counterweight can be connected to the input shaft to rotate as a single unit with it. This connection can be achieved by manufacturing the shaft and counterweight as a monolithic piece, or by fixing the counterweight to the shaft by means such as press fitting, welding, or bolting. The expression "countering the unbalanced dynamic force" describes the function of the counterweight, which is to generate a dynamic inertial force that, at any point during the rotation of the input shaft, is approximately equal in magnitude and opposite in direction to the unbalanced dynamic force generated by the cycloidal disk. The superposition of these two opposing forces results in a substantial reduction or cancellation of the net dynamic force exerted on the main bearings of the input shaft, leading to smoother operation with less vibration. Structurally, this can be achieved by placing the center of mass of the counterweight diametrically opposite the center of mass of the combined cycloidal disk and its eccentric bearing, relative to the main axis of rotation of the input shaft. For example, if the eccentric section of the input shaft is offset from the main axis in a first radial direction, the counterweight can be positioned on the shaft so that its mass lies predominantly in a second radial direction, 180 degrees opposite the first radial direction. The mass of the counterweight can be selected such that the product of its mass and its radial distance from the axis of rotation substantially matches the product of the mass of the cycloidal disk assembly and its eccentric displacement. The superposition of these two opposing forces can result in a substantial cancellation of the net dynamic force exerted on the main input shaft bearings, leading to smoother operation and less vibration. This actuator provides a novel configuration that results in a compact and lightweight structure with a high torque output capacity. The counterweight can be rigidly connected to the eccentric input shaft and is positioned to generate an inertial force that is substantially equal in magnitude and opposite in direction to the unbalanced dynamic force generated by the single orbiting cycloidal disk. The superposition of these opposing forces substantially cancels the net dynamic force exerted on the input shaft bearings. This configuration allows the actuator to operate with low vibration levels, similar to a conventional twin-disc system, but more complex. The operational coupling of the components provides a single-stage, high-ratio speed reduction. The rotation of the eccentric input shaft imparts an orbital motion to the cycloidal disk. This orbital motion, restricted by external fixed elements, causes the cycloidal disk to rotate at a significantly lower speed than the input shaft. This kinematic arrangement allows the actuator to produce high output torque from a high-speed, low-torque motor within a compact physical volume. The output element is driven directly by the slow rotation of the cycloidal disk, providing the actuator's final high-torque motion. The entire power transmission path is designed to be robust and capable of withstanding the shock loads that occur during the robot's dynamic operation, ensuring the actuator's mechanical reliability. The counterweight can be in the shape of an arch. This actuator may comprise a single cycloidal disc. The term "single cycloidal disc," as used here, specifies that the actuator's cycloidal reducer comprises exactly one cycloidal disc to perform the speed reduction function. This contrasts with conventional cycloidal reducer designs, which typically employ a pair of identical cycloidal discs, often arranged in a 180-degree offset configuration to achieve inherent dynamic balance. Under this design, all eccentric motion and torque transmission are handled by this single disc. The use of a single cycloidal disc significantly simplifies the gearbox's mechanical structure. By definition, this configuration eliminates the second disc, the second disc bearing, and the need for a double eccentric input shaft required in a conventional double-disc system. This reduction in the total number of components directly translates into a decrease in the actuator's overall axial length and weight, contributing to a higher power-to-weight ratio. Furthermore, the use of a single cycloidal disc concentrates all functional loads on a single set of components. The single disc, its associated bearing, and the output drive elements are configured to withstand the entire torque and all contact forces of the gearbox. This concentrated load path, combined with the dynamic balance provided by the counterweight, allows for a design that is both structurally simple and operationally robust, without the alignment complexities inherent in dual-disc systems. The use of a single cycloidal disc represents a preferred embodiment of the invention. However, it should be understood that the principle of dynamic equilibrium with a counterweight is also applicable to other gearbox configurations that are not inherently balanced. For example, an alternative embodiment of the actuator could comprise two or more cycloidal discs that are not arranged in a dynamically balanced configuration with a 180-degree phase shift. Such a system, even with multiple discs, would still generate an unbalanced net dynamic force during operation and would therefore benefit from the inclusion of a counterweight on the input shaft to ensure smooth, low-vibration performance. Nevertheless, the single-disc embodiment provides the greatest degree of mechanical simplification and compactness. The robot leg actuator may further comprise a fixed housing with a plurality of fixed annular pins configured to engage with a lobed outer profile of the cycloidal disc. The term "fixed housing," as used herein, refers to a stationary, non-rotating structural component of the actuator that serves as a frame of reference for the mechanism. It may contain and support internal components, including input shaft bearings and fixed ring pins, and may provide mounting points for attaching the actuator to the robot structure. The term "fixed annular pins" refers to a series of cylindrical posts or rollers that can be rigidly mounted within the fixed housing. These pins can be arranged in a circular pattern, concentric with the main shaft of the actuator, and their purpose is to provide a stationary race or track against which the cycloidal disc can roll. The term "lobed outer profile" describes the non-circular, multi-lobed shape of the outer perimeter of the cycloidal disk. This profile can be a high-order curvilinear shape, such as an epitrochoid, and can be precisely engineered to mesh with the fixed ring pins. The number of lobes in this profile can be directly related to the gearbox's speed reduction ratio, typically being one less than the number of fixed ring pins. The expression "configured to mesh" describes the functional and geometric relationship between the lobed outer profile of the cycloidal disk and the fixed ring pins. This means that the components are sized and positioned so that, as the cycloidal disk performs its orbital motion, its lobed profile makes continuous contact with the fixed ring pins, one after the other. This engagement is what limits the disk's movement and translates the high-speed orbital motion into low-speed rotation. The inclusion of a fixed housing with a plurality of fixed ring pins provides the stationary reference frame against which the cycloidal disc operates. This structural arrangement is useful for the operation of the cycloidal reducer. The fixed housing provides the necessary rigidity and precise positioning of the ring pins, ensuring accurate and repeatable engagement with the lobed profile of the disc. The interaction between the lobed outer profile of the cycloidal disc and the array of fixed annular pins is the central mechanism for speed reduction. As the eccentric input shaft drives the disc in its orbital path, the engagement with the fixed pins forces the disc to roll along the inner circumference of the pin circle. Since the number of disc lobes differs from the number of pins, this rolling motion results in a slow, high-torque rotation of the disc relative to the fixed housing. This arrangement allows for a high reduction ratio to be achieved in a single, compact stage. Each of the fixed ring pins may comprise a plastic bearing, which may be configured to provide a low-friction rolling surface. The term "plastic bearing," as used herein, can refer to a sleeve- or roller-like component made of a polymeric material, which is positioned around each of the fixed ring pins. This component can act as an interface between the fixed pin and the moving cycloidal disc. The plastic material can be selected for its specific mechanical properties, such as a low coefficient of friction, high wear resistance, and good dimensional stability. Examples of such materials include polyamides (PA), polyoxymethylene (POM), and other engineering polymers. The term "low-friction rolling surface" describes the primary function of the plastic bearing. The outer surface of the plastic bearing is the surface that comes into direct contact with the lobed outer profile of the cycloidal disc. By being able to rotate freely around its respective fixed ring pin, the bearing transforms what would be a sliding friction contact into a much more efficient rolling friction contact. The term "low friction" means that the coefficient of friction at this interface is significantly reduced compared to direct sliding contact between metal and metal or between metal and plastic. Using a plastic bearing on each fixed ring pin substantially reduces internal friction in the cycloidal reducer. As the cycloidal disc rolls against the pins, these bearings rotate, minimizing friction losses that would otherwise generate heat and reduce the overall efficiency of the actuator. This reduction in internal friction directly improves the actuator's recoil capability. Recoil capability, or the ability of the output to be moved by an external force, is critical for legged robots to achieve compliant control, absorb ground impacts, and regenerate energy. By minimizing static and dynamic friction within the gearbox, the plastic bearings allow external forces applied to the robot leg to be transmitted through the gearbox to the motor with minimal resistance, enabling more responsive force control and more natural, elastic limb behavior. The output element can be actuated by the cycloidal disc through a plurality of output pins. The term "output pins" refers to a set of cylindrical posts or shafts that may be part of the reducer's output element or rigidly connected to it. These pins may be arranged in a circular pattern, for example, concentric with the output element's axis, and may be configured to pass through corresponding holes or openings in the cycloidal disc. They serve as a direct mechanical interface for transferring the low-speed, high-torque rotation from the cycloidal disc to the final output element of the actuator. Using multiple output pins to transfer motion from the cycloidal disk to the output element provides a robust and distributed means of torque transmission. The load can be shared among all the output pins, reducing stress on any individual pin and on the corresponding contact surfaces within the cycloidal disk. This load-distribution arrangement increases the overall torque capacity and shock resistance of the gearbox. Furthermore, the multi-pin coupling ensures stable, backlash-free motion transfer, which is useful for precise position control in robotic applications. The output bushings can be arranged to provide an interface between the cycloidal disk and the output pins. The term "output bushings" refers to hollow cylindrical components that can be positioned within the through-holes of the cycloidal disc. The bushings can be arranged with a defined clearance within these holes, allowing them a limited degree of movement, or "float," relative to the cycloidal disc. This floating arrangement may be necessary due to the distinct kinematics of the components. The output pins, being rigidly connected to the output element, can undergo pure rotational motion around the actuator's central axis. In contrast, the cycloidal disk does not undergo pure rotation but follows a complex orbital motion (or cycloidal motion) in which its center continuously orbits around the actuator's central axis. Therefore, the function of the floating output bushings is to act as a kinematic coupling that accommodates the difference between these two motions. As the cycloidal disk orbits and rotates simultaneously at low speed, the bushings slide and can shift within their respective holes in the disk to compensate for the discrepancy between the orbital path of the holes and the purely circular path of the output pins they contain. This decoupling of the slight translational component of the disk's motion from the output pins allows for smooth, continuous, and jam-free transmission of torque from the orbital disk to the rotating output element. Each outlet cap may comprise an outer metal layer and an inner plastic layer. The term "outer metal sleeve" refers to the outer part of the bushing, made of metal or metal alloy, such as steel, aluminum, or bronze (or alloys thereof). Its outer surface may be configured to slide with low friction against the inner diameter of the cycloidal disc holes to facilitate the necessary compensating movements. The outer metal sleeve may be configured to slide against the edge of the corresponding holes in the cycloidal disc. The term "inner plastic sleeve" refers to the internal lining of the bushing, made, for example, from a polymer (low friction and wear-resistant) or other material, such as polyoxymethylene (POM), polyamide, or polyetheretherketone (PEEK). The polymer may be fiber-reinforced and / or may, alternatively or additionally, include polytetrafluoroethylene (PTFE). This choice or material may lead to increased self-lubrication of the bushing. The inner sleeve may be made from a low-friction metal, such as bronze. Alternatively, the inner sleeve may be made from a relatively soft material (compared to the outer sleeve). Typically, the outer sleeve is made of a relatively hard material (compared to the inner sleeve). The inner sleeve can be made of a low-friction material. Alternatively, a single-layer sleeve comprising a hard outer surface and a low-friction inner surface, such as a metal alloy (e.g., bronze) or hardened metal (e.g., steel), can be used. Its inner bore provides a rotating bearing surface for the output pins. This bi-material construction allows the bushing to effectively manage the two distinct types of relative motion that occur simultaneously at the interface. The metal-on-metal (or metal-on-composite) contact between the outer sleeve and the cycloidal disc is optimized for the slight translational slip required for kinematic compensation. At the same time, the plastic-on-metal contact between the inner sleeve and the output pins can be optimized for the continuous rotary motion of torque transfer. This separation of functions within a single component results in a highly efficient, durable, and precise output mechanism. The robot leg actuator may further comprise an output bearing configured to rotationally support the output element with respect to a fixed actuator housing, said output bearing comprising an outer race and an inner race. The term "output bearing," as used here, refers to a mechanical element that restricts the relative motion between two parts (for example, here the rotating output element and the stationary housing) to the desired rotational motion. Its primary function is to reduce friction between these moving and stationary parts and to support the loads acting upon them. The bearing can be of any suitable type, including, but not limited to, rolling element bearings (such as ball bearings or roller bearings) or plain bearings (such as bushings or sliding bearings). The expression "rotational support of the output element" describes the primary function of the output bearing. It means that the bearing provides a stable mechanical restraint that allows the output element to rotate freely around a defined axis relative to the fixed housing, while resisting forces that would cause it to move in any other direction. These resisted forces can include translational forces (both radial and axial) and tilting moments that could cause the output element to tilt or wobble. The term "outer ring" refers to the outer ring of the output bearing, which may be fixed. The term "inner ring" refers to the inner ring of the output bearing, which may be rotating. One function of this output bearing is its integration into the overall structure of the robotic joint. It can be specifically designed not only to support the internal loads generated by the reduction mechanism but also to withstand the significant external structural loads applied to the robotic limb during operation, such as the weight of the leg and impact forces from the ground. This eliminates the need for separate, additional bearings in the robot's joint assembly, leading to a significant reduction in the number of parts, overall weight, and mechanical complexity of the robot. This integrated design can be useful for creating very compact and lightweight robotic joints. The outer race of the output bearing can be fixedly connected to a fixed actuator housing, and the inner race of the output bearing can be fixedly connected to the output element. The output bearing can be a double-row angular contact bearing or a crossed roller bearing configured to support axial and radial loads and to counteract tilting moments. The term "double-row angular contact bearing" refers to a type of rolling element bearing comprising two rows of balls. The contact lines between the balls and the raceways can be angled with respect to the radial plane. This configuration allows the bearing to accommodate combined loads, meaning it can support both radial (perpendicular to the axis of rotation) and axial (parallel to the axis of rotation) loads in both directions. The term "crossed roller bearing" refers to a type of rolling element bearing in which the cylindrical rollers are arranged in an alternating, crisscrossed pattern, for example, with each roller oriented at 90 degrees to the adjacent one. The rollers may contact raceways with V-grooves. This arrangement allows a single bearing to support significant loads from all directions: radial loads, axial loads, and moments. The expression "configured to withstand axial and radial loads and counteract tilting moments" describes the structural capability of these specific types of bearings. Due to their internal geometry (angular contacts of balls or crossed rollers), they are inherently capable of withstanding forces applied simultaneously from multiple directions. "Tipping moments" (or moment loads) are rotational forces that tend to tilt or rock one bearing race relative to the other. The ability to counteract these moments provides a high degree of rotational stiffness and stability to the joint. Selecting a double-row angular contact bearing or a crossed roller bearing for the output bearing is a design choice that enables complete integration of the actuator into the robot's joint. Unlike simpler radial bearings, these specific types provide the high rigidity and multi-axis load capacity required to serve as the primary structural bearing for a robotic limb. Their ability to counteract tipping moments is especially important in a single-disc cycloidal reducer, as they provide the stability needed to resist tilting forces generated by asymmetric internal loads. This allows the actuator to function as a compact, self-contained, and structurally robust joint module. The robot leg actuator may further comprise a brushless motor configured to drive the eccentric input shaft. The term "brushless motor," also known as a brushless direct current (BLDC) motor or electronically commutated motor, refers to a type of synchronous electric motor that uses permanent magnets in the rotor and electromagnets in the stator. Unlike a conventional brushed motor, commutation (the process of changing the current in the motor windings to generate motion) is achieved electronically by a motor controller, rather than by mechanical brushes. The motor controller may use sensors, such as Hall effect sensors, or sensorless techniques to determine the rotor position and energize the stator windings in the correct sequence to produce rotation. The inclusion of a brushless motor as the actuator's main motor provides several key technical advantages. The absence of mechanical brushes eliminates a primary source of wear, friction, and electrical noise, resulting in significantly greater reliability, a longer lifespan, and reduced maintenance requirements compared to brushed motors. Furthermore, brushless motors offer a higher torque-to-weight ratio and greater efficiency, allowing for a more powerful and compact actuator for a given input power. Electronic commutation also enables highly precise control of the motor's speed and torque, which is beneficial for the advanced control strategies employed in legged robot locomotion. The robot leg actuator may further comprise a motor controller, a first encoder for measuring the rotational position of the motor, and a second encoder for measuring the rotational position of the output element. The term "motor controller" refers to an electronic circuit, for example, implemented on a printed circuit board (PCB), that governs the performance of a brushless motor. It takes high-level commands (such as the desired speed or torque) and converts them into precisely timed, high-power electrical signals that are applied to the motor's stator windings. To do this, it may process feedback from sensors, such as the first encoder, to perform electronic switching and may implement closed-loop control algorithms for torque, speed, and position. The term "encoder" refers to a sensor or transducer that converts mechanical motion (in this case, rotation) into a digital or analog electrical signal. This signal provides information about the angular position, speed, and / or direction of a rotating shaft. The encoder can be of any suitable type, including magnetic encoders (which use Hall effect or magnetoresistive sensors to detect the position of a magnet) or optical encoders (which use a light source and a photodetector to read a patterned disk). The term "rotational position of the output element" refers to the absolute or relative angular orientation of the actuator's final output element after the speed reduction has occurred. This is the actual position of the robotic joint itself and is the primary variable that must be controlled to achieve precise limb movement. The inclusion of an integrated motor controller and a dual encoder system (a first encoder for the motor and a second encoder for the output) enables advanced, high-performance control of the actuator. The first encoder provides high-resolution information about the motor's status, which is useful for the motor controller to perform smooth and efficient electronic switching and implement wide-bandwidth torque and speed control loops. The second encoder, by directly measuring the rotational position of the output element, provides a true and accurate measurement of the robotic joint's position. This allows the system to compensate for any nonlinearity, backlash, or torsional flex within the reduction mechanism itself. A control system utilizing feedback from both encoders can achieve superior position accuracy and can also detect discrepancies between the motor and output positions. This information can be used to estimate external torques, detect collisions, and implement compatible control strategies without requiring a separate, dedicated torque sensor. This dual-encoder architecture creates a highly transparent and responsive robotic actuator. A magnetic shielding plate can be placed between a magnet of the first encoder and a magnet of the second encoder to prevent magnetic interference. The term "magnetic shielding plate" refers to a plate or layer of material with high magnetic permeability, such as a mu-metal alloy or a sheet of soft iron. Its purpose is to deflect or block magnetic field lines, thereby isolating a specific region from external magnetic fields. The term "magnetic interference," also known as magnetic crosstalk, refers to the undesirable effect whereby the magnetic field generated by one component (e.g., the magnet of the first encoder) influences the operation of an adjacent magnetic sensor (e.g., the sensor of the second encoder). This interference can disrupt sensor readings, resulting in inaccurate position measurements and degraded control performance. The placement of a magnetic shielding plate between the magnets of the first and second encoders is a feature that allows for a compact, colocalized arrangement of the dual encoder system. By positioning the two encoder systems very close together, the overall size of the actuator's electronic assembly is minimized. However, this proximity can lead to significant magnetic interference, as the strong field of one magnet would distort the field measured by the other sensor. The magnetic shielding plate solves this problem by positioning itself in the path of these stray magnetic fields. The plate's high-permeability material provides a low-reluctance path for the magnetic flux, effectively capturing and diverting the field lines from the interfering magnet away from the sensitive sensor area of ​​the other encoder. This isolation ensures that each encoder operates solely with the magnetic field of its own corresponding magnet, resulting in clean, accurate, and reliable position signals for both the motor and the output. This allows the actuator to benefit from the control advantages of a dual-encoder system without sacrificing design compactness. This document also proposes a robotic leg comprising at least one actuator as mentioned above. The term "robotic leg," as used herein, refers to an articulated limb of a legged robot, which may be responsible for providing support and propulsion to the robot. A robotic leg is typically a kinematic chain composed of two or more rigid or semi-rigid links, such as a thigh and a leg, connected by one or more actuated joints, such as a hip joint and a knee joint. The leg is designed to interact with the ground or other surfaces to enable locomotion, such as walking, running, or climbing. By constructing a robotic leg using one or more actuators according to the invention, the entire limb benefits from the actuator's characteristics. The use of these compact and lightweight actuators allows for the design of a leg with low mass and low inertia, especially at its distal end. This reduces the energy required to balance the leg during locomotion, enabling higher speeds and greater overall robot efficiency. Furthermore, the high output torque and robust, shock-resistant nature of the actuators allow the leg to withstand the significant forces generated during dynamic movements and impacts with the ground, ensuring the reliability of the entire robotic system. This document also proposes a legged robot comprising a body and a plurality of robotic legs, as previously mentioned. The term "legged robot" refers to a mobile robot that uses articulated limbs, or legs, for locomotion, instead of wheels or tracks. This category of robots includes, among others, bipedal (two-legged) robots, quadrupedal (four-legged) robots, and hexapod (six-legged) robots. These robots can be designed to traverse complex, uneven, or unstructured terrain that would be inaccessible to wheeled vehicles. The term "body," as used herein, refers to the central chassis or torso of the legged robot. The body may serve as the main structure to which the robot's multiple legs can be attached. It may also house the main power source (e.g., batteries), the main computer and control systems, and any mission-specific payload. This robot with legs can be a quadruped robot comprising four of these robotic legs. This robot with legs can be a bipedal robot comprising two of these robotic legs. This robot with legs can be a hexapod robot comprising six of these robotic legs. This document also refers to a robotic leg actuator, comprising: - a speed reducer having an input shaft and an output element; - the speed reducer comprises at least one cycloidal disc configured to be driven in an orbital motion by the input shaft; - a plurality of output pins connected to the output element and passing through corresponding holes in at least one disk; - a plurality of output bushings, each bushing disposed within one of said holes to provide an interface between the at least one disk and one of said output pins, in which each output bushing is configured to provide a wear-resistant outer surface to cooperate with the cycloidal disc and a low-friction inner surface to cooperate with the output pin. The outlet bushing may include: - an outer sleeve made of a first material that provides said outer surface resistant to wear; and - an inner sleeve made of a second material, different from the first material, which provides said low-friction inner surface, the inner sleeve being arranged coaxially within the outer sleeve. The first material can be a metallic material and the second material can be a plastic material. The second material can be selected from the group comprising bronze and polytetrafluoroethylene (PTFE). The outlet bushing can be a single-layer bushing made from a monolithic material that provides both the wear-resistant outer surface and the low-friction inner surface. The monolithic material can be a bronze alloy or a surface-hardened steel. This document also refers to a robot leg actuator, comprising: - a motor with a rotor; - a motor-driven speed reducer having an output element; - a first magnetic encoder configured to measure the rotational position of the rotor, the first magnetic encoder comprising a first magnet and a first sensor; - a second magnetic encoder configured to measure the rotational position of the output element, the second magnetic encoder comprising a second magnet and a second sensor; - a magnetic shielding plate, for example, made of a high permeability material, and arranged between the first magnet and the second magnet, said plate is configured to prevent magnetic interference between the first and second magnetic encoders. Brief description of the drawings Other features, details, and advantages will become apparent from the detailed description below and from the analysis of the accompanying drawings, in which: - Figure 1 is a perspective view of a quadruped robot incorporating the actuators according to an embodiment of the present invention. - Figure 2 is a perspective view of a single-leg robotic assembly, showing the arrangement of three actuators. - Figure 3 is a cross-sectional view of a complete set of actuators, showing the integration of its main components. - Figure 4 is an exploded view showing the main subassemblies of the actuator. - Figure 5 is a perspective view of the mounted actuator. - Figure 6 is a cross-sectional view showing the integration of the motor and controller units. - Figure 7 is an exploded view of the motor controller unit and its electronic components. - Figure 8 is a cross-sectional view of the cycloidal reduction unit. - Figure 9 is an exploded perspective view of the cycloidal reduction unit. - Figure 10 is a front view of the cycloidal disk assembly, illustrating the coupling of the cycloidal disk with the fixed ring pins and the output pins. - Figure 11 is a cross-sectional view of a custom bi-material outlet bushing. - Figure 12 is a schematic block diagram illustrating the main functional components of the cycloidal reducer. Detailed description of the drawings Figures 1 to 10 illustrate a non-limiting embodiment of this document. Throughout the figures, the same reference numbers are used to designate identical or corresponding components. First, Figure 1 shows a legged robot 1, which in this non-limiting embodiment is a quadruped robot 1. The robot comprises a main body or torso 2 and four articulated legs 4. The body 2 is generally rectangular or parallelepiped in shape and serves as the central chassis for the robot 1. It is configured to house essential components, such as the main power supply (e.g., batteries) and control computers, and may carry various sensors to perceive its environment. These sensors may include, among others, front-facing stereo cameras 6 for depth perception, panoramic cameras for a 360-degree field of view, and an inertial measurement unit (IMU) to measure orientation and acceleration. Each leg 4 is an articulated kinematic chain designed to provide support and propulsion. As illustrated in the detailed view of a single leg in Figure 2, each leg 4 comprises several distinct links. A thigh link 8 is connected to the robot body 2 via a hip joint assembly 10. A lower leg link, or shin 12, is connected to a distal end 14 of the thigh link 8 by a knee joint 16. At the end of the shin 12 is the foot 18, which is the terminal portion of the leg 4 designed to make contact with the ground. Links such as the calf 12 may have a lightweight, lattice, or truss structure to provide high rigidity with minimal mass. To provide the degrees of freedom necessary for complex locomotion, each leg 4 is actuated by a series of three independent actuators 20, 22, 24, which are the subject of the present invention. These actuators 20, 22, 24 are arranged as follows: - A first actuator 20 is mounted on the hip and is oriented to control the abduction and adduction of leg 4. This corresponds to a rotation of leg 4 around the balance axis (X-axis), as defined in Figure 1.1, allowing leg 4 to move laterally with respect to body 2. A second actuator 22 is also mounted on the hip, adjacent to the first actuator 20. It is configured to control the flexion and extension of the entire leg 4 from the hip. This corresponds to a rotation of the thigh link 8 with respect to the body 2, around the pitch axis (Y-axis), allowing the thigh link 8 to rock forward and backward. - A third actuator 24 is also mounted on the hip, adjacent to the second actuator 20. This actuator 24 is configured to control the flexion and extension of the calf 12 with respect to the thigh 8. This also corresponds to a rotation about a tilt axis (Y-axis), effectively bending and straightening the leg at the knee 16. The coordinated action of these three actuators 20, 22, 24 gives leg 4 three degrees of freedom, allowing the robot's control system to accurately position foot 18 in three-dimensional space and apply controlled forces to the ground to achieve various types of gait, from walking to running and jumping. Each actuator 20, 22, 24 (the actuator will be referred to as 20 in the following description) is composed of three main subassemblies arranged coaxially: - an engine control unit 26, which houses the electronic control components. - a motor unit 28, which contains the frameless, brushless motor. - a cycloidal reduction unit 30, which provides speed reduction and torque amplification. These three modules 26, 28, 30 are designed to be mounted into a single, compact, sealed unit, as shown in the assembled view of Figure 5 and the cross-sectional view of Figure 3. The motor control unit 26 is located at one end of the actuator 20. As detailed in the exploded view of Figure 7, it comprises a main motor control board 32 and a dual encoder system. This entire electronic assembly is protected by an actuator cover 34, which also serves as a heat sink with its external fins. The motor unit 28 is located in the central part of the actuator 20. It is a frameless, brushless motor, meaning that its main components, the stator 36 and rotor 38, are integrated directly into the structure of the actuator 20 without their own housing. The stator 36 is fixed inside the motor housing 40, while the rotor 38 is mounted on a shaft assembly 42 that connects to the gearbox 30. The cycloidal reduction unit 30 is located at the opposite end from the actuator 20 to the controller 26. This unit 30 is responsible for converting the high-speed, low-torque rotation of the motor into a low-speed, high-torque output. The main structural component of this unit 30 is the reducer housing 44. The actuator output is provided by a gearbox output shaft 46. The output shaft 46 has a mounting face 48 at its outer end. This mounting face 48 is provided with a pattern of threaded holes 50 (Figures 3 and 8), which are configured to receive fasteners such as screws. This interface 48, 50 is designed for the direct mechanical attachment of an adjacent robot limb or component. For example, a flange on the thigh link 8 of a robot can be screwed directly to this mounting face 48, creating a rigid connection that transmits torque and motion from the actuator to the limb. The fixed structure of the actuator 20 is mainly composed of the reducer housing 44, the motor housing 40, and the actuator cover 34. These components are rigidly fixed together to form a single sealed outer housing. The gearbox housing 44 is located at the output end of the actuator 20 and serves as the main structural mounting point. It is generally a cylindrical component with a flange 52 or mounting holes 54 (Figure 5) for attachment to the robot frame. Internally, the gearbox housing 44 contains a circular arrangement of fixed annular pins 56 that are securely mounted in the housing 44, for example, by press fit. These pins 56 protrude axially and are fitted with freely rotating plastic bearings 58, which provide a low-friction surface for a cycloidal disc 60. The housing 44 also contains the seat 62 for the outer race of the main output bearing 64. The motor housing 40 is rigidly attached to the gearbox housing 44. It is a hollow, finned cylinder designed to enclose and support the stator 36 of the brushless motor. The external fins 66 increase the surface area of ​​the housing 40 to improve heat dissipation from the motor during operation. The actuator cover 34 is fixed to the end of the motor housing 40 opposite the reducer 30. It encloses and protects the motor control unit 26. The cover 34 can also contribute to heat dissipation and provides mounting points for the external power and data connectors 68 (Figure 4). The rotor 38 of the brushless motor is the main rotating element. It is rigidly mounted on the shaft assembly 42. The shaft assembly 42 is integrated or rigidly coupled to an eccentric input shaft 70 of the cycloidal reducer 30. The entire input shaft assembly 42, 70 is supported by bearings 72 and rotates as a single unit when the motor is activated. A counterweight 74 is fixed to the shaft assembly 42, located diametrically opposite the eccentric area or journal 76 of the input shaft 70 to ensure dynamic balance, as shown in Figure 10. The counterweight 74 is arc-shaped and is fixed to the shaft assembly by means of pins 78. The counterweight 74 extends at an angle of between 0 and 180°, preferably between 30 and 150°, for example at an angle of 83°. The cycloidal disc 60 is mounted so that it rotates freely on the eccentric input shaft 70, by means of a bearing 80, for example a ball bearing. As the input shaft 70 rotates at high speed, the eccentric journal forces the cycloidal disc 60 into an orbital motion. The disc 60 comprises a lobed outer profile 82 (Figure 10) that rolls against the inner track formed by the series of fixed plastic bearings 58 mounted on the fixed pins 56. This restricted rolling motion forces the cycloidal disc 60 to rotate slowly about its own center in the opposite direction to the rotation of the input shaft 70. This slow, high-torque rotation of the cycloidal disc 60 is transferred to the output shaft of the final reducer 46. This is achieved by means of a set of output pins 84, which are an integral part of or fixed to the output shaft 46 (Figures 9 and 10). These pins 84 pass through the corresponding through holes 86 of the cycloidal disc 60. The interface between the holes 86 of the disc 60 and the output pins 84 is mediated by output bushings 88 that are mounted to rotate freely around the pins 84. As the disc 60 orbits, the output shaft 46 undergoes pure rotation. A retainer 88 (Figures 8 and 9) can be used to secure the bushings 88. Each bushing 88 is a bi-material component (Figure 11) designed to optimize both structural integrity and low-friction performance. It comprises an outer metal sleeve 88A, which forms the outer body of the bushing. This sleeve is typically made of a hard metal alloy. It also comprises an inner plastic sleeve 88B, which is arranged coaxially and bonded to the outer metal sleeve. This inner sleeve is made of a high-performance, low-friction polymer. The inner bore of this plastic sleeve 88B provides the bearing surface against which the output pin 84 rotates. The entire rotary output assembly, including the output shaft 46, is supported by the integrated main output bearing 64, which bears all the structural loads of the joint. Finally, a reducing cover 90 and a dynamic seal 92, held by a retainer 94, close the assembly, ensuring its protection and axial restraint of the output bearing 64. The actuator is a fully integrated mechatronic module, comprising its own integrated control electronics housed within the motor control unit 26. This integration eliminates the need for complex external wiring to a central robot controller, increasing reliability and simplifying the overall robot design. The motor control unit 26 is located at the proximal end of the actuator 20, protected by the actuator cover 34. The core of this unit is the motor control board 32. This board contains the necessary power electronics, such as a three-phase inverter bridge, to drive the stator windings 36, as well as a microprocessor or digital signal processor to execute advanced control algorithms. One feature of the control system is a dual encoder architecture for precise and robust motion control. This system provides high-resolution information on both the motor status and the final position of the output shaft. As best seen in the exploded view of Figure 7, a first encoder 96 is configured to measure the rotational position of the motor's rotor 38. This encoder comprises a first magnet 98 (Figures 3 and 6), which is mounted at one end of the shaft assembly 42, and a corresponding magnetic sensor (e.g., a Hall effect or magnetoresistive sensor) located on a first readout plate 100. The motor controller 32 uses this high-frequency feedback to perform smooth electronic switching and to implement high-bandwidth torque and speed control loops. A second encoder 102 is configured to measure the rotational position of the reducer's output shaft 46 via a central auxiliary shaft 104. End 106 of this auxiliary shaft is fixed to the output shaft 46, and the other end 108 supports a second magnet 110. This auxiliary shaft 104 is housed within the input shaft 13. The second encoder 102 comprises this second magnet 110 and a second reading plate 112. This provides an absolute measurement of the joint angle. By measuring the final output position, this encoder 102 can be used to increase the accuracy of the output positioning. To allow this dual encoder system to be integrated into the compact volume of the motor control unit 26, the two encoder assemblies 96, 102 are arranged very close to each other. To prevent magnetic interference between them, a magnetic shielding plate 114 is placed between the first magnet 98 and the second magnet 110. This shielding 114, made of a high-permeability material, isolates the two sensors from each other, ensuring that each provides a clean and reliable signal. Figure 12 illustrates an example of a robotic leg actuator 20 comprising a cycloidal reducer 30 according to this document, the cycloidal reducer 30 comprising: - an eccentric input shaft 70; - a cycloidal disk 60 operatively coupled to the eccentric input shaft 70, such that the rotation of the input shaft 70 imparts an orbital motion to the cycloidal disk 60, said orbital motion generating an unbalanced dynamic force; and - an output element 46 operatively coupled to the cycloidal disk 60; The actuator 20 is characterized in that the cycloidal reducer 30 further comprises a counterweight 74 rotatably connected to the eccentric input shaft 70 and configured to counteract the unbalanced dynamic force generated by the orbital motion of the cycloidal disc 60.

Claims

1. A robotic leg actuator (20), comprising: a speed reducer (30) having an input shaft (70) and an output element (46); the speed reducer (30) comprising at least one cycloidal disk (60) configured to be driven in an orbital motion by the input shaft (70); a plurality of output pins (84) connected to the output element (46) and passing through corresponding holes (86) in the at least one disk (60); and a plurality of output bushings (88), each bushing being disposed within one of said holes (86) to provide an interface between the at least one cycloidal disk (60) and one of said output pins (84), wherein each output bushing (88) is configured to provide a wear-resistant outer surface to cooperate with the cycloidal disk (60) and a low-friction inner surface to cooperate with the output pin (84). 2.The actuator according to the preceding claim, wherein the output sleeve (88) comprises: - an outer sleeve (88A) made of a first material providing said wear-resistant outer surface; and - an inner sleeve (88B) made of a second material, different from the first material, providing said low-friction inner surface, the inner sleeve (88B) being arranged coaxially within the outer sleeve (88A).

3. The actuator according to the preceding claim, wherein the first material is a metallic material and the second material is a plastic material.

4. The robotic leg actuator (20) according to any of the preceding claims, wherein the output sleeves (88) are arranged with a defined clearance within said holes (86) in the cycloidal disc (60). 5.The robotic leg actuator (20) according to any of the preceding claims, wherein said speed reducer (30) is a cycloidal reducer comprising a single cycloidal disc (60).

6. The robotic leg actuator (20) according to any of the preceding claims, wherein the cycloidal reducer (30) further comprises a counterweight (74) rotatably connected to the eccentric input shaft (70) and configured to counteract the unbalanced dynamic force generated by the orbital motion of the cycloidal disc (60).

7. The robotic leg actuator (20) according to any of the preceding claims, further comprising a fixed housing (44) with a plurality of fixed annular pins (56) configured to engage with a lobed outer profile (82) of the cycloidal disc (60). 8.The robotic leg actuator (20) according to any of the preceding claims, wherein the output element (46) is driven by the cycloidal disc (60) via a plurality of output pins (84).

9. The robotic leg actuator (20) according to any of the preceding claims, further comprising an output bearing (64) configured to rotatably support the output element (46) with respect to a fixed housing (44) of the actuator (20), said output bearing (64) comprising an outer race and an inner race.

10. The robotic leg actuator (20) according to the preceding claim, wherein the outer race of the output bearing (64) is fixedly connected to a fixed housing of the actuator (20), and wherein the inner race of the output bearing is fixedly connected to the output element (46). 11.The robotic leg actuator (20) according to claim 9 or 10, wherein the output bearing (64) is a double-row angular contact bearing or a crossed roller bearing configured to support both axial and radial loads and to counteract tipping moments.

12. The robotic leg actuator (20) according to the preceding claim, further comprising a motor (28) configured to drive the eccentric input shaft (70), a motor controller (32), a first encoder (96) for measuring a rotational position of said motor (28), and a second encoder (102) for measuring a rotational position of the output element (46).

13. The robotic leg actuator (20) according to the preceding claim, wherein a magnetic shielding plate (114) is disposed between a magnet (98) of the first encoder (96) and a magnet (110) of the second encoder (102) to prevent magnetic interference. 14.A robotic leg (4) comprising at least one actuator (20) according to any of the preceding claims.

15. A legged robot (1) comprising a body (2) and a plurality of robotic legs (4) according to the preceding claim.