S3E: A Multi-Robot Multimodal Dataset for Collaborative SLAM

TL;DR

S3E is introduced, an expansive multimodal dataset that not only surpasses previous efforts in scale, scene diversity, and data intricacy but also provides a thorough analysis and benchmarks for both collaborative and individual SLAM methodologies.

Abstract

The burgeoning demand for collaborative robotic systems to execute complex tasks collectively has intensified the research community's focus on advancing simultaneous localization and mapping (SLAM) in a cooperative context. Despite this interest, the scalability and diversity of existing datasets for collaborative trajectories remain limited, especially in scenarios with constrained perspectives where the generalization capabilities of Collaborative SLAM (C-SLAM) are critical for the feasibility of multi-agent missions. Addressing this gap, we introduce S3E, an expansive multimodal dataset. Captured by a fleet of unmanned ground vehicles traversing four distinct collaborative trajectory paradigms, S3E encompasses 13 outdoor and 5 indoor sequences. These sequences feature meticulously synchronized and spatially calibrated data streams, including 360-degree LiDAR point cloud, high-resolution stereo imagery, high-frequency inertial measurement units (IMU), and Ultra-wideband (UWB) relative observations. Our dataset not only surpasses previous efforts in scale, scene diversity, and data intricacy but also provides a thorough analysis and benchmarks for both collaborative and individual SLAM methodologies. For access to the dataset and the latest information, please visit our repository at https://pengyu-team.github.io/S3E.

Figures (6)

• Figure 1: Mobile Platform Sensor Layout and Coordinate Systems. The left part details the sensor locations and the coordinate frames that define their spatial orientation relative to the platform. In the right part, our mobile platforms are available in two versions, each designed for different operational requirements.
• Figure 2: Illustration of Trajectory Paradigms for C-SLAM. The distinct trajectory paradigms adopted by three agents, designated as Alpha, Bob, and Carol, to demonstrate different interaction and information exchange patterns in a multi-agent SLAM context. (a)Concentric Circles. (b) Intersecting Circles. (c) Intersection Curve. (d) Rays
• Figure 3: Visualization of Outdoor Trajectory in the S3E Dataset. The outdoor trajectories captured in the S3E dataset by three tele-operated mobile platforms, designated as Alpha ($\alpha$), Bob ($\beta$), and Carol ($\gamma$). The trajectories are distinctly annotated with Orange, Purple, and Cyan. The annotations $\bigstar$, $\blacktriangle$, and $\mdblkcircle$ indicate the specific positions where data was recorded in each sequence. The left portion of the figure presents the synchronized data capture at the annotated points, demonstrating the collaborative data collection process.
• Figure 4: S3E Dataset Organizational Structure.
• Figure 5: Map in Laboratory_1 with CoRLIO.