How to Set Up a Mobile Robotics Lab from Scratch: Equipment, Budget, and Mistakes to Avoid

TL;DR: A functional mobile robotics lab setup needs four pillars: a reliable UGV platform, a ROS 2 development workstation, a structured test area with motion reference, and a networking layer that survives multi-robot traffic. Expect a baseline budget between approximately 15,000 and 40,000 EUR for a small research group of 2 to 4 people. The most common mistakes are underestimating power and charging logistics, skipping version control for ROS 2 workspaces, and buying sensors before defining the research question.

Why a Structured Mobile Robotics Lab Setup Matters

A mobile robotics lab is not a collection of robots on a shelf. It is a reproducible environment where you can test navigation stacks, sensor fusion, and autonomy algorithms under controlled conditions, then export the same software to field trials. Without structure, you spend more time debugging cables and Wi-Fi than running experiments.

This guide walks through the equipment, budget ranges, and spatial organization needed for a working lab around ROS 2 and platforms such as the Leo Rover. It targets research groups, university courses, and small integrator teams building UGV-based systems.

What Equipment Belongs in a Mobile Robotics Lab

The equipment list breaks into five functional groups. Each group serves a specific stage of the development cycle, from code editing to outdoor validation.

Mobile robot platforms

Start with one or two platforms that match your research question. For indoor SLAM and navigation work, a differential-drive or 4-wheel skid-steer UGV in the 5 to 10 kg range is enough. For outdoor field robotics, you need IP-rated chassis, brushed or brushless motors with sufficient torque, and a payload bay for additional sensors. The Leo Rover, for example, offers a 5 kg payload, IP66 rating, and a ROS 2 compatible stack out of the box, which removes weeks of low-level driver work.

Compute and development workstations

Each developer needs a Linux workstation, preferably Ubuntu 22.04, 24.04, or 26.04 LTS to match ROS 2 Humble, Jazzy, or Lyrical where supported by your tools and robot vendor. Minimum practical specs: 32 GB RAM, an 8-core CPU, and a discrete GPU if you plan to run perception models or Gazebo / Isaac Sim simulations. Onboard the robot, an NVIDIA Jetson Orin Nano or a Raspberry Pi 5 with a Coral accelerator covers most use cases.

Sensors

Buy sensors after you know what you want to measure. A defensible starter kit includes:

• A 2D LiDAR such as the RPLIDAR A2 or SLAMTEC C1 for indoor mapping
• A depth camera such as the Intel RealSense D435i or Orbbec Gemini for visual SLAM and obstacle detection
• An IMU, optionally with magnetometer, for dead reckoning and orientation
• A u-blox ZED-F9P RTK GNSS module if you plan outdoor work with centimeter-level accuracy under RTK corrections and open-sky conditions

Networking and infrastructure

ROS 2 uses DDS for transport, and DDS is sensitive to network conditions. A dedicated Wi-Fi 6 access point on a separate VLAN, plus a wired gigabit switch for stationary nodes, prevents most multi-robot communication issues. Add a managed router that supports static IP leases for every robot.

Power, tools, and safety

This is the layer most often forgotten. You need a charging station with at least two parallel chargers per robot type, spare battery packs sized for at least 2x to 3x expected session runtime, anti-static mats, a soldering station, a labeled cable bin, and an emergency stop procedure documented on the wall. For outdoor work, add a hard-shell transport case and a weatherproof tarp.

What a Realistic Budget Looks Like

The table below shows orientation ranges for a small research lab supporting 2 to 4 engineers. Prices are approximate, in EUR, based on publicly listed vendor pricing as of 2024 to 2026.

Budget table for a starter mobile robotics lab

• Mobile robot platform (1 to 2 units): 4,000 to 12,000 EUR. A ROS 2 ready UGV such as the Leo Rover sits in the lower half of this range for a single-unit configuration.
• Development workstations (2 to 4 units): 4,000 to 10,000 EUR. Linux laptops or desktops with GPUs.
• Sensors (LiDAR, depth camera, IMU, GNSS): 1,500 to 6,000 EUR depending on accuracy class.
• Networking (Wi-Fi 6 AP, switch, router): 400 to 1,200 EUR.
• Power and charging (batteries, chargers, PSU): 500 to 2,000 EUR.
• Tools and workbench (soldering, oscilloscope, hand tools): 800 to 3,000 EUR.
• Test arena materials (mats, fiducials, obstacles): 300 to 2,000 EUR.
• Motion capture or fiducial reference system (optional): 2,000 to 30,000+ EUR. A low-cost AprilTag setup is at the bottom; OptiTrack or Vicon at the top.

Total entry-level configuration lands around 15,000 EUR. A mid-range lab with motion capture and two robots reaches 35,000 to 40,000 EUR. University-grade installations with four robots and full Vicon coverage cross 80,000 EUR.

How to Organize the Physical Space

Lab layout determines how fast you can iterate. A workable mobile robotics lab needs three zones with clear boundaries.

Workstation zone

Desks with dual monitors, wired Ethernet drops, and reachable power. Keep at least one workstation directly adjacent to the test arena so engineers can monitor live telemetry without walking across the room.

Test arena

A minimum of 5 by 5 meters of clear floor for indoor UGV testing. Walls or curtains in matte, non-reflective material reduce LiDAR ghosting. Mark a coordinate origin on the floor with tape and place AprilTags at known positions for ground-truth localization. Avoid glossy floors. Polished concrete can reflect IR and degrade active depth cameras.

Maintenance and storage zone

Separate the dirty work, soldering, drilling, gluing, from the clean robots and workstations. Store batteries in a fire-rated LiPo bag or metal cabinet. Label every cable. A whiteboard with the current state of each robot, charged, in use, broken, saves hours per week.

What Mistakes to Avoid When Building a Robotics Lab

The following issues come up repeatedly in robotics education and research labs. Each is avoidable with one decision made early.

• Buying hardware before defining the research question. A 360-degree 3D LiDAR is useless if your project is line-following. Write the experiment specification first, then size the sensors.
• Ignoring the ROS 2 distribution lifecycle. Pick one LTS distribution per project phase and stick to it. Mixing Humble, Jazzy, and Lyrical nodes across robots can create dependency, message, QoS, and middleware issues that are painful to debug.
• No version control for workspaces. Every robot should boot from a Git-tracked configuration. Use a tool such as vcstool or a Docker image registry to pin dependencies.
• Underestimating power logistics. One charger per robot is often not enough. You will lose half a working day waiting for batteries. Plan enough spare batteries and 2x to 3x charging capacity.
• Mixing test and development networks. Student laptops on the same Wi-Fi as robots flood DDS discovery traffic. Separate VLANs or a dedicated SSID solves this.
• No reproducible test scenarios. If you cannot run the same navigation test twice, you cannot publish results. Define start poses, obstacle layouts, and success criteria in writing.
• Skipping outdoor validation. Indoor SLAM looks great until sunlight hits the depth camera. If your application is field robotics, budget for outdoor sessions from week one.

How to Choose Between Education and Research Configurations

The same hardware can serve both purposes, but the workflow differs. In an education context, you optimize for student throughput: identical robots, locked-down software images, and short reset cycles between lab sessions. In a research context, you optimize for experimental flexibility: root access on every robot, customizable payload mounts, and detailed logging of every sensor stream.

One platform family across both use cases reduces support overhead. The Leo Rover, for instance, runs the same ROS 2 stack whether used in a 12-week undergraduate course or in a multi-year field robotics PhD, which means lab staff maintain one set of skills instead of three.

FAQ: Mobile Robotics Lab Setup

What is the minimum budget to start a mobile robotics lab?

A two-person lab with one ROS 2 capable UGV, two workstations, basic sensors, and networking starts around 15,000 EUR. This excludes motion capture and outdoor test infrastructure.

Which ROS 2 distribution should a new lab adopt?

Pick the most recent LTS release supported by your robot vendor and your simulation tools. As of 2026, ROS 2 Jazzy is the safest default for many new labs, Humble remains practical for existing Ubuntu 22.04 stacks, and Lyrical is relevant where your vendor and tools already support Ubuntu 26.04.

Do you need motion capture for a research lab?

Only if your research requires millimeter to sub-centimeter ground truth, for example multi-robot coordination or aggressive maneuvering. For navigation, SLAM, and most autonomy research, well-calibrated AprilTag-based localization at 1 to 3 cm accuracy can be enough.

Can one robot platform serve both indoor and outdoor experiments?

Yes, if it has an IP-rated chassis and enough payload margin. Platforms in the 5 to 10 kg class with IP54 or higher, such as the Leo Rover, handle both lab floors and gravel paths without hardware swaps.

How much space is needed for an indoor test arena?

5 by 5 meters is the practical minimum for a single UGV running navigation tests. For multi-robot work or larger platforms, plan 8 by 8 meters or more.

What is the most underestimated cost item?

Power and charging logistics. Spare batteries, parallel chargers, and a safe storage cabinet often add 1,000 to 2,000 EUR that was not in the original plan.

How long does it take to commission a new lab?

From purchase order to first reproducible experiment, expect 6 to 12 weeks. Hardware delivery, network setup, software image preparation, and writing the first test protocols each take longer than predicted.

Where to Go Next

If you are planning a mobile robotics lab for a university group, an R&D team, or a field robotics startup, start by writing the first three experiments you want to run, then size the equipment against those specifications. Explore the research applications page for examples of how Leo Rover deployments are configured across academic and industrial labs, and review available platforms and accessories in the Fictionlab shop.