R, a commonly used language for statistical analyses, offers robust statistical capabilities making it a good fit for a myriad of engineering applications. To explore basic functionality, this article explores the UR3 CobotOps dataset, which was contributed to the UC Irvine Machine Learning Repository in 2024.

One of the biggest selling points of R is that it is open source. Its strengths in engineering lie in its robust statistical analysis tools, strong data visualization capabilities, and a rich ecosystem of specialized packages. The language also has some limitations in engineering contexts. It may not be as efficient for large-scale numerical computations, real-time data processing, or complex simulations compared to languages like MATLAB or Python. Additionally, R’s syntax can be less intuitive for engineers more familiar with traditional programming languages. Note, the author’s coding skills are developing. As a burgeoning R programmer, this analysis relied on the assistance of online tutorials as well as with coding assistance from large language models, most notably Claude Sonnet 3.5. 

Dataset overview

The UR3 CobotOps Dataset is a comprehensive collection of multi-dimensional time-series data from the UR3 cobot. For context, the UR3 from Universal Robots is a compact, lightweight collaborative robot arm designed to automate repetitive tasks in various industrial settings. The dataset, which includes 20 features, captures data on operational parameters including electrical currents, temperatures, and speeds across its six joints (J0-J5). It also has data on gripper current, operation cycle count, protective stops, and grip losses. This data, collected via MODBUS and RTDE protocols, provides a detailed operational snapshot of the cobot.

Relevance to engineering

This UCI dataset has several potential engineering applications, including:

1. Fault detection: By analyzing patterns in joint currents and temperatures, engineers can identify potential faults before they lead to failures.
2. Predictive maintenance: Understanding the relationships between operational parameters can help in developing predictive maintenance models.
3. Operational optimization: Insights from this data can lead to improved control algorithms and more efficient task allocation.

R in action

Now, let’s explore how R can help us explore the UCI data on the UR3 cobot. We’ll focus on three key areas: time series analysis, distribution analysis, and correlation analysis.

1. Time series analysis of joint currents

One of R’s strengths is its ability to handle and visualize time series data effectively. Using ggplot2, we created a time series plot of joint currents. Robot joints, or axes, are crucial components in cobots like the UR3, enabling rotational or translational movement between connected segments. The UR3 features six joints — base, shoulder, elbow, and three wrist joints — each providing a single degree of freedom. These joints, with the help of motors and sensors, enable the cobot to perform an array of precise motions, and generally simulate the versatility of a human arm.

The first is a chart is a display of current in A through the first joint, here labelled “Joint 0′ over time. For context, Joint 0 is the base joint — the foundational component of the robot arm — and the time interval between consecutive timestamps in the dataset roughly 1 second.

• The image shows the dynamic behavior of each joint over time, with clear spikes.
• There are distinct operational patterns, with periods of high activity interspersed by relative calm.

In general, this visualization type can help spot anomalies, shed light on operational patterns, and potentially detect early signs of wear or malfunction in specific joints.

2. Distribution of joint temperatures

To understand the thermal characteristics of the cobot, we created a boxplot of joint temperatures:

In the dataset, Joints 4 and 5 (J4 and J5) operate at higher temperatures than the others. The manual for the cobot describes Joints 4 and 5, as part of the wrist. They are thus involved in more complex and precise movements which are correlated with increased motor activity and friction​.

In general, a box plot can assist with thermal management, predictive maintenance, and ensure the cobot operates within safe temperature ranges.

3. Correlation Analysis

To understand the relationships between different operational parameters, a correlation heatmap can be helpful:

This visualization offers several insights:

• There are strong positive correlations between temperatures of adjacent joints.
• Some joint currents show negative correlations with their speeds.
• The tool current shows correlations with various joint parameters.

These visualizations demonstrate R’s capability in extracting insights from complex engineering data. Such analyses form the foundation for more advanced applications like predictive maintenance models, fault detection algorithms, and performance optimization strategies.

Additional Visualizations

Here are some additional visualizations that provide further insights into the UR3 cobot’s operation:

This scatter plot shows the relationship between current and speed for Joint 0 (the base joint), which can shed light on the energy consumption patterns of the joint under different operational conditions.

The above histogram displays the distribution of tool current, capturing the typical power consumption of the end effector and potentially identifying different operational modes. For context, the end effector is the device at the end of the robotic arm that interacts with the environment, such as a gripper, welder, or other tools used for specific tasks.

This bar plot above shows the correlations between various parameters and the tool current, helping to identify which factors most strongly influence the end effector’s power consumption. This visualization reveals significant current fluctuations across all joints, with joints J0, J1, and J5 showing higher amplitudes. As collaborative robots continue to evolve and integrate into diverse R&D environments, the ability to extract and interpret such detailed operational data — increasingly freely available via open source licenses — could open up new opportunities for understanding robotic capabilities.