--- title: "Robotic Household Activities Dataset" description: "This droid dataset includes recordings of cleaning, laundry folding, dishwashing activities, combining head-mounted video with detailed robotic data streams from seven 9-axis IMU units placed…" url: "https://unidata.pro/datasets/robotic-household-activities/" date_modified: "2026-05-19T14:52:55+03:00" language: "en-US" --- This droid dataset includes recordings of cleaning, laundry folding, dishwashing activities, combining head-mounted video with detailed robotic data streams from seven 9-axis IMU units placed on the chest, upper arms, forearms, wrists. The robot dataset captures data from 3-axis accelerometers, 3-axis gyroscope, and 3-axis magnetometer sensors, using onboard sensor-fusion to generate orientation quaternions for motion tracking and robotic learning. ## Dataset Structure ### The Numbers Section **Numbered list:** - **Number:** 1,000 — **Text:** Hours ### Tooltips Section **Tooltip items:** - **Name:** Computer Vision - **Name:** Machine Learning - **Name:** Human Activity Recognition - **Name:** Robot Learning - **Name:** Motion Analysis ### Dataset Information **Table with data:** | Characteristic | Data | | --- | --- | | Description | 1,000 hours of multimodal recordings of cleaning, laundry folding, and dishwashing activities | | Data types | Video (head-mounted), 9-axis IMU streams | | Tasks | Human Activity Recognition, Motion Analysis, Multimodal Learning, Action Segmentation | | Hours of recordings | 1,000 | | IMU placements | Chest, Upper arms, Forearms, Wrists | | Sensor outputs | 3-axis accelerometer, 3-axis gyroscope, 3-axis magnetometer, Orientation quaternions | | Labeling | Metadata (counter, duration, task, date, video size, sensor size, recording version) | **Media Slider:** - **Video on Slayder:** - **Video on Slayder:** **Link to the sample:** [Download sample](https://drive.google.com/drive/folders/1FxrNajp1IKjtbAFKwfrXi56tfWNFlGzl) ### Technical Specifications **Table with data:** | Characteristic | Data | | --- | --- | | Video extension | MP4 | | Video source | Head-mounted camera | | Extension of labeling file | JSON | | IMU specification | 7 units, 9-axis each | | IMU data types | Linear acceleration, Angular velocity, Magnetic field orientation | | Orientation format | Quaternions (from onboard sensor-fusion) | **Source and data collection methodology:** Source and collection methodology. Data was collected by a partner of Unidata. ### Dataset Use Cases - Slider **Industry Cards:** - **Industry:** Robotics & Automation — **Title:** Learning Household Manipulation Skills — **Text:** This dataset supports robotic learning for everyday manipulation tasks such as cleaning, laundry folding, and dishwashing. With 1,000 hours of synchronized head-mounted video and 9-axis IMU sensor data, it enables model training on real-world motions, trajectories, and object interactions, helping robots learn complex sequences in realistic home environments. - **Industry:** Artificial Intelligence Research — **Title:** Training Foundation Models with Multimodal Data — **Text:** The robot activities dataset is well suited for training foundation models that combine vision, sensor data, and human demonstrations. By capturing detailed sensor data from real robots and humans, researchers can improve model performance in action segmentation, motion understanding, and multimodal learning methods across large-scale robot datasets. - **Industry:** Humanoid Robotics — **Title:** Human Motion Imitation and Skill Transfer — **Text:** For humanoid robotics, this dataset provides valuable training data to study how human movements translate to robot arms and full-body systems. The inclusion of precise orientation quaternions and IMU streams helps models learn smooth motions, grasp strategies, and coordinated actions needed for human-like task execution. - **Industry:** Robotic Systems Engineering — **Title:** Benchmarking Real-World Robotic Performance — **Text:** Engineers can use this robot dataset to benchmark robotic systems under realistic scenarios involving varied objects and tasks. The structured metadata, high-quality video, and sensor recordings allow consistent evaluation of robotic manipulation, making it useful for comparing learning methods and validating performance in real-world applications. ### Fact **FAQs Heading:** FAQs **List of Questions:** - **Question:** What is included in this robotic dataset? — **Answer:** The dataset includes 1,000 hours of multimodal recordings featuring cleaning, laundry folding, and dishwashing tasks. It consists of head-mounted video and synchronized sensor data streams collected from real humans performing household activities. - **Question:** What types of annotations are provided in the dataset? — **Answer:** Annotations include structured metadata such as task type, recording duration, timestamps, recording version, and sensor file sizes. These labels enable precise alignment between video, sensor data, and manipulation tasks for model training. - **Question:** Can I request a sample of the dataset before purchasing or downloading it? — **Answer:** Yes. Unidata provides free samples so you can evaluate the video quality, sensor data structure, and labeling format. This allows researchers to assess whether the robotic data supports their robotic systems and learning methods before full access. - **Question:** What are the sources of data for Unidata robotic datasets? — **Answer:** This robot activities dataset was collected by a verified partner of Unidata. Data collection followed controlled protocols to ensure consistent, high-quality robot and human demonstration recordings. - **Question:** What sensor data does the dataset contain? — **Answer:** The dataset contains data from seven 9-axis IMU units placed on the chest, upper arms, forearms, and wrists. Each IMU provides linear acceleration, angular velocity, magnetic field orientation, and orientation quaternions generated through onboard sensor fusion. - **Question:** What makes this dataset different from standard activity recognition datasets? — **Answer:** Unlike traditional activity datasets that rely only on video or wearable sensors, this robot dataset combines synchronized first-person video, seven 9-axis IMU streams, orientation quaternions, and structured metadata. This rich multimodal data makes it particularly valuable for robot learning, human activity recognition, motion analysis, and next-generation Physical AI applications. - **Question:** How are Unidata datasets licensed? — **Answer:** Unidata follows a dual-licensing model. Free samples are available for testing and evaluation, while full robotic datasets are provided exclusively through purchase. - **Question:** Do Unidata datasets comply with GDPR and data privacy regulations? — **Answer:** Yes. All datasets are curated in compliance with GDPR and applicable data protection laws. Data is collected from legally permissible sources to ensure ethical use in robotics and machine learning research. - **Question:** How are Unidata datasets stored and managed? — **Answer:** Unidata stores all datasets securely on AWS cloud infrastructure with compliance to ISO 27001 and ISO 27701 standards. This ensures high availability, scalability, and secure handling of robotic data and sensor streams. - **Question:** How long does it take to receive the dataset after purchase? — **Answer:** Once you submit a request, Unidata will contact you to confirm requirements and complete documentation. After signing and payment, the dataset is delivered within 3–10 days. ## List of Parameters - **Title:** Tasks — **Description:** Human Activity Recognition, motion analysis, action segmentation - **Title:** Sensor data — **Description:** 3-axis accelerometer, gyroscope, magnetometer + orientation quaternions - **Title:** Video — **Description:** Head-mounted MP4, JSON metadata - **Title:** Hours — **Description:** 1,000 - **Title:** Use — **Description:** Robotic learning, foundation models, humanoid systems [Full list of this site's AI-readable pages](https://unidata.pro/llms.txt)