Apply Advances in Sensing, Connectivity, and Motion Control Devices for Smarter Fixed-in-Place Robots

Mounted (fixed-in-place) robotic systems, often referred to as multi-axis robots, are designed to deliver high-precision, high-performance motion within a defined workspace. These systems form the backbone of modern manufacturing and automation cells, where repeatability, speed, and payload capacity are critical.

Common examples include collaborative robots (cobots), articulated robot arms, selective compliance articulated robot arms (SCARAs), and delta (parallel-link) mechanisms, as well as computer numerical control (CNC) and gantry-style machines. Depending on application requirements, such robots can be mounted on rails, walls, ceilings, floors, or integrated directly into production machinery, enabling flexible deployment for assembly, material handling, packaging, inspection, and machining processes.

By combining advanced drive electronics, precision sensors, and real-time control architectures, these mounted robotic platforms provide the reliability, versatility, functionality, and accuracy essential for smart, connected manufacturing environments. However, to maximize the benefits and performance of these systems, designers must understand and apply the latest advances in motion detection, position and area sensing, motion control, and connectivity.

This article provides a brief examination of the design requirements of advanced robots. It then introduces example solutions and related evaluation kits from Analog Devices that designers can use to implement these systems.

Design requirements of advanced robots

Advanced mounted robots (Figure 1) have two distinctions compared to mobile robots: they operate within a relatively static and known overall environment, and they are not constrained by battery power. However, they are expected to operate with speed, precision, repeatability, and accuracy despite changing circumstances. For example, they may need to pick up packages of varying sizes, shapes, weights, orientations, and positions, and place them in a precise location on a moving belt. To do this, these robots must autonomously assess and dynamically adapt to the situation while maintaining awareness of their settings and surroundings.

Figure 1: The well-known, widely seen fixed-in-place robot now offers extreme precision, flexibility, and adaptability. (Image source: Analog Devices Inc.)

These requirements necessitate the careful integration of precise motion control for end effectors, time-of-flight (ToF) imaging for ambient awareness, inertial measurement units (IMUs) for motion sensing, and Gigabit Multimedia Serial Link (GMSL) for reliable, high-speed communication.

1: Motion control for end effector grippers: Robot grippers function like hands or clamps, opening and closing on demand. They must use the appropriate amount of force to maintain a firm grip without damaging the payload. Doing so requires the motor drive to carefully modulate the motor for precise, consistent, and smooth operation. The drive should also have low mass and be compact due to weight and space constraints.

One suitable solution for such a controller is the TMCM-1617 single-axis servo drive (Figure 2). Weighing 24 g and measuring 36.8 mm × 26.8 m × 11.1 mm, this three-phase brushless-DC (BLDC) motor drive can supply up to 18 A RMS and operates from an 8 V to 24 V supply.

Figure 2: The lightweight and compact TMCM-1617 servo drive provides a complete 8 V to 24 V, 18 A BLDC motor control. (Image source: Analog Devices Inc.)

The TMCM-1617 supports incremental encoders and digital Hall-effect sensors for position feedback, thus enhancing its precision and repeatability under varying loads. For connectivity, it features CAN, RS-485, and EtherCAT bus options.

To quickly evaluate and tune the TMCM-1617 and its algorithms, Analog Devices offers the TMCM-1617-GRIP-REF Gripper Reference Design. This open-source hardware reference design is tailored for precise control of 24 V BLDC motors used in robotic grippers. It provides precise field-oriented control (FOC), ensuring minimum torque ripple and enabling efficient, high-performance motor control. The pre-configured software stack streamlines the initial setup process, reducing time to market.

2: ToF sensors: Designers have two basic choices for ensuring that the robot is fully aware of its surroundings and any objects in its operating zone: use a ToF sensing arrangement or use one or more video cameras. Each offers relative advantages and shortcomings.

In general, ToF cameras are preferred for depth sensing and offer high-accuracy distance measurement. However, they typically have lower spatial resolution than conventional video cameras and can be affected by ambient light and reflective surfaces. Standard video cameras, on the other hand, provide high-resolution images and are versatile for various applications, but extracting depth information requires more complex processing and multiple cameras.

For many robotic applications, the benefits of ToF-based imaging are significant. A ToF-based sensing subsystem, however, requires the careful integration of many electro-optical components, including a matched LED light source, lenses, optical filters, and an imager. Selecting and assembling these components requires extensive electrical, mechanical, and optical expertise.

To minimize these difficulties, Analog Devices offers the ADTF3175 ToF module (Figure 3). This complete unit features a 1 megapixel (MP) CMOS indirect ToF imager. It also integrates a lens and a 940 nm optical band-pass filter for the imager, an infrared-illumination source containing optics, a laser diode, a laser-diode driver and photodetector, flash memory, and power regulators to generate local supply voltages.

Figure 3: The ADTF3175 module includes all the necessary electronic, mechanical, and optical elements for a complete ToF subsystem. (Image source: Analog Devices Inc.)

The image-cloud data output of the 1024 × 1024 pixel ADTF3175 sensor (with a 75° × 75° field of view (FOV)) is sent to the host system over a four-lane Mobile Industry Processor Interface (MIPI) Camera Serial Interface 2 (CSI-2) operating at 1.5 gigabits per second (Gbits/s) per lane. The module’s programming and operation are controlled through a four-wire serial peripheral interface (SPI) and an I²C interface. The depth range is 0.4 to 4 meters (m) with a depth accuracy of ±5 mm across the full depth range.

An associated ADSD3500 depth-image signal processor converts the megapixel-resolution raw data from the ADTF3175 to produce final radial depth, active brightness (AB), and confidence data frames. This ensures low latency with high frame rates, allowing the camera to precisely capture fast-moving objects and enabling the robot to make timely decisions and provide accurate analyses in dynamically changing industrial settings.

To facilitate the setup and implementation of the module, Analog Devices offers the EVAL-ADTF3175D-NXZ 3D ToF sensor evaluation kit (Figure 4). This open-source kit includes the ADTF3175 module, a third-party numerical processing unit for embedded artificial intelligence (AI) and machine learning (ML) applications, a camera interface board, an interposer adapter board, and a tripod.

Figure 4: The EVAL-ADTF3175D-NXZ evaluation kit provides the necessary processing, connectors, and tripod to facilitate design-in of the ADTF3175 ToF sensor. (Image source: Analog Devices Inc.)

3: IMU: As the robot end effector (grip) is free to move anywhere within its prescribed three-dimensional zone, it is critical to know both its location and orientation in that space. One way to do this is to use encoders at each joint, then combine all their outputs by using coordinate transformation and matrix equations. However, this requires multiple multi-axis encoders and adds computational complexity.

An attractive alternative is to use a six degrees-of-freedom (6 DoF) IMU that combines a three-axis (triaxial) accelerometer and a three-axis gyroscope. The ADIS16500 miniature microelectromechanical systems (MEMS) IMU (Figure 5, left) provides this capability in a tiny 15 × 15 × 5 mm package with an SPI output. The associated ADIS16500/PCBZ evaluation board (Figure 5, right) measures 33.25 mm × 30.75 mm. This board serves primarily as a breakout board to facilitate wiring connections via a 16-pin (2 × 8), 2 mm pitch connector to the comprehensive EVAL-ADIS2Z evaluation system.

Figure 5: The high-level block diagram of the ADIS16500 (left) only hints at the internal integration and sophistication of this 6-DoF IMU; the associated ADIS16500/PCBZ breakout board (right) serves primarily as a physical-connection interface to the EVAL-ADIS2Z evaluation system. (Image source: Analog Devices Inc.)

The digital gyroscopes feature a ±2,000° per second (˚/s) dynamic range, while the digital accelerometers offer a ±392 m/s² dynamic range. Each inertial sensor in the ADIS16500 line includes signal conditioning that optimizes dynamic performance.

In addition, as gyroscopes and accelerometers have unique inherent error sources, factory calibration is used to characterize each sensor in terms of sensitivity, bias, alignment, linear acceleration (gyroscope bias), and point of percussion (accelerometer location). As a result, each sensor has dynamic compensation formulas that yield extremely accurate sensor measurements over a broad set of conditions.

4: GMSL: There are important considerations when merging all these functional blocks in a robotic arm: they must be linked together, and they, particularly the ToF module, generate a large amount of time-critical data. The GMSL interface addresses these situations. Originally developed for automotive use, GMSL has been adopted by applications such as robotics, as it supports the required high data rates on a single cable.

For example, the MAX96724 deserializer, in an 8 × 8 mm TQFN package, converts four GMSL 2/1 inputs to 1, 2, or 4 MIPI D-PHY or C-PHY paths (Figure 6). This 6 Gbit/s, four-input, two-output device allows simultaneous transmit bidirectional transmissions over 50 ohm (Ω) coaxial or 100 Ω shielded twisted-pair (STP) cables. The device supports up to four remotely located sensors.

Figure 6: The MAX96724 deserializer converts four GMSL 2/1 inputs to 1, 2, or 4 MIPI D-PHY or C-PHY paths. (Image source: Analog Devices Inc.)

Each GMSL2 serial link operates at a fixed rate of 3 or 6 Gbits/s in the forward direction and 187.5 Mbits/s in the reverse direction. The link can also automatically adapt the forward-path receiver characteristics to compensate for the insertion loss and return loss characteristics of the channel; these losses are largely determined by cables, connectors, temperature effects, and the properties of the printed circuit board (pc board). The MAX96724 supports both aggregation and replication of video data, enabling streams from multiple remotely located sensors to be combined.

These are complicated devices to set up and use. Analog Devices made the task easier with the MAX96724-BAK-EVK# evaluation kit (Figure 7). This kit provides a proven design and a reliable platform for evaluating the MAX96724 devices using standard FAKRA coaxial cables (a rugged cable/connector assembly used in automotive and other applications) or a MATE-AX cable (a miniaturized version of FAKRA cables).

Figure 7: The MAX96724-BAK-EVK# evaluation kit is a valuable tool for implementing a design based on the highly sophisticated MAX96724. (Image source: Analog Devices Inc.)

The kit includes a simple-to-use Windows 10-compatible (or higher) graphical user interface (GUI) to exercise device features.

Conclusion

State-of-the-art mounted robotic systems require careful integration of multiple technologies to perform with the needed speed, precision, and flexibility. Using diverse technologies, including advanced servo control, ToF imaging, and IMUs, all connected using a GMSL, they can implement and integrate the needed functionality. Analog Devices offers the necessary components along with evaluation units to accelerate the design-in process and minimize risk and uncertainty.

Related Content

1. Webinar: “Robotics with GMSL: High-Speed Vision and Perception in Real Time”
2. Analog Devices, “Time of Flight System Design—Part 1: System Overview”