In recent years, sensor technology has advanced rapidly worldwide, and robots increasingly rely on sensors for control operations and data feedback. Sensors have become indispensable “sensory organs” for robots. During high-complexity tasks, robots require precise coordination across various stages, and faults in these stages are often difficult to detect. Therefore, the development of sensor technology has garnered significant attention from researchers, and the demand for signal acquisition and data analysis of sensors has increased notably. However, existing methods for sensor signal acquisition suffer from poor efficiency and low real-time performance, which are critical issues that need addressing in the industry.
A six-axis force sensor can simultaneously detect all information of forces in space, obtaining force components Fx, Fy, Fz and moment components Mx, My, Mz along the X, Y, and Z directions. This capability significantly enhances the precision of robotic operations and finds important applications in contact operations, assembly, grinding, and bimanual coordination tasks. Despite progress in decoupling characteristics, linearity, and coupling factors of six-axis force sensors, such as those based on Stewart structures or wireless designs, there remains a gap in efficient and stable signal detection methods. Current approaches often lack accuracy and reliability, necessitating improved solutions.
To address these challenges, we designed a signal acquisition system for six-axis force sensors based on the TCP/IP protocol. This system leverages a data acquisition card for low-noise amplification and multi-channel signal acquisition, coupled with a custom-developed upper-computer software programmed in C# on the Visual Studio 2019 platform. The software facilitates system configuration, data processing, storage, and real-time curve display, achieving low-cost, high-efficiency, and strong human-computer interaction. Ethernet communication following the TCP/IP protocol enables seamless data transmission by simply setting IP and port addresses on the client and server sides. The acquired data is displayed as real-time curves on the upper computer, allowing users to monitor robotic operations promptly and prevent potential failures. The system is portable, and through software simulation and comparative experiments, we have verified its stability, feasibility, high real-time performance, and ease of operation.
The overall design of the system comprises hardware and software modules. The hardware includes the six-axis force sensor, a low-noise amplification module, a data acquisition card, and an upper computer. The software module handles system settings, data processing, data storage, and real-time curve display. Figure 1 illustrates the system block diagram, where force signals are converted into electrical signals, conditioned by the low-noise amplification module, acquired by the data acquisition card, and transmitted via Ethernet to the upper computer for analysis and display.
In the hardware design, we selected the M8128 data acquisition instrument from Junde Technology for low-noise amplification and data acquisition. This instrument integrates key modules, supports six-channel analog input, and employs Ethernet TCP/IP protocol communication, RS232, and CAN bus communication. Its working principle involves a main control module, communication module, and power supply module, providing compact size, portability, high cost-effectiveness, precision, and adaptability to complex environments. Compared to traditional methods using PLCs as intermediaries, our system directly acquires signals from the six-axis force sensor via the data acquisition card, simplifying hardware configuration. The upper computer is a commercial microcomputer running our custom software, supporting network communication and capable of handling large data volumes in confined spaces.
Key technical parameters of the M8128 data acquisition instrument are summarized in Table 1.
| Parameter | Value |
|---|---|
| Number of Channels | 6 |
| Maximum Sampling Rate | 2 kHz |
| Resolution | 24-bit |
| Trigger Source | AC/DC |
The software design is programmed in C# due to its simplicity, stability, and security, widely used in application development. The system software structure includes client and server components, with the server handling network communication, data storage, and parsing, while the database stores data for client access. The software features data processing, storage, and real-time curve display. Ethernet communication allows easy IP and port configuration for real-time monitoring of force and moment changes in the six-axis force sensor. The client interface is user-friendly, displaying multi-channel real-time curves for intuitive observation of robotic motion states.
The program flow begins with system initialization, setting timer intervals, defining array structures, and initializing real-time curves for force and moment components with distinct colors. After configuring server and client IP and ports, the TCP server connection is established. Acquired data undergoes base conversion for processing, and force and moment calculations are performed using the following formulas:
$$F_{xyz}(N) = F_{xyz} \times \text{Scale Factor}[w1] / \text{CPF}$$
$$M_{xyz}(N \cdot m) = M_{xyz} \times \text{Scale Factor}[w2] / \text{CPT}$$
Here, w1 takes values 0, 1, 2 for force components, and w2 takes 3, 4, 5 for moment components. The results are displayed in real-time on the computer screen. Socket communication based on TCP/IP protocol ensures reliable data transmission, as TCP is stream-oriented and avoids data loss compared to UDP. The network connection involves server listening, client request, and connection confirmation.
For system validation, we conducted software simulation and comparative experiments. Using Zonalan debugging tools, we simulated server-side data transmission to the client in TCP server mode, with IP set to 127.0.0.1 and port 49151. By sending 24-byte conversion parameter commands and 16-byte parameters, we observed real-time curve changes matching calculated values. For instance, inputting specific hex data yielded force and moment results as shown in Table 2, confirming system effectiveness and real-time performance.
| Sequence | Fx (N) | Fy (N) | Fz (N) | Mx (N·m) | My (N·m) | Mz (N·m) |
|---|---|---|---|---|---|---|
| 1 | 0.102 | 0.966 | 0.688 | 0.707 | 0.199 | 0.096 |
| 2 | 0.113 | 0.962 | 0.705 | 0.122 | 0.214 | 0.100 |
| 3 | 0.120 | 0.770 | 0.668 | 0.079 | 0.236 | 0.063 |
In comparative testing, we evaluated the system against a wireless signal acquisition system. As shown in Figure 10, both systems achieved accuracy around 80%, but our system demonstrated superior stability with minimal fluctuation over multiple tests, indicating higher reliability. This highlights the advantages of our six-axis force sensor signal acquisition system in precision and consistency.
The system offers several benefits over traditional force gauges, including real-time display of force and moment curves, immediate user feedback, low cost, and simplicity. It provides a foundation for subsequent robotic fault analysis and motion state monitoring. Future work could focus on enhancing software algorithms for advanced data analytics and integrating with cloud-based platforms for remote monitoring.
In conclusion, our design of a six-axis force sensor signal acquisition system based on TCP/IP protocol effectively addresses existing gaps in signal detection. The hardware and software integration ensures efficient, real-time data acquisition and processing, making it suitable for various industrial applications. The system’s portability, stability, and user-friendly interface further contribute to its practicality in modern robotic systems.