iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
A robot that remembers where it has been needs more than a single “memory” component. At minimum, it needs a sensor that observes the environment, a continuous estimate of its own motion, software and compute that match observations to places, and storage if the map must survive a reboot. A 2D lidar with robot odometry is one documented route to an indoor 2D map, but lidar is not mandatory. RGB-D cameras, stereo cameras, IMUs, visual odometry, and other sensors can fill the same roles when the chosen software supports them.
The right bill of materials depends on what spatial memory means for your project: a metric map for navigation, repeat localization, named destinations, or semantic recall of rooms and objects. The sections below separate those requirements so you do not buy hardware that your mapping stack cannot use.
What does a robot spatial-memory system actually have to do?
Spatial memory is a pipeline, not a part number. The robot observes surfaces or features, estimates short-term motion, builds or retrieves a map, and stores enough information to recognize places later.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →- Observation: A lidar, RGB-D camera, stereo camera, or another environmental sensor measures geometry or visual features.
- Motion estimation: Wheel odometry, an IMU, visual odometry, lidar odometry, or a fused combination estimates how the robot moved between observations.
- Mapping and localization: SLAM or a localization system relates observations to a coordinate frame. Nav2 describes the global
map → odomtransform and the locally continuousodom → base_linktransform as separate responsibilities. - Persistence: A disk database or saved map lets the system resume after shutdown. Active working memory and long-term storage are different resources.
- Application memory: Room names, object identities, landmarks, and task-specific labels require data above a basic geometric occupancy grid.
A stationary robot, or one that only processes recorded sensor files, can omit mobile-base hardware and live power constraints. A mobile robot navigating around people cannot: it needs current obstacle observations even when it has a previously saved map.
Popular on Amazon
- 【90 Degree Full Function USB C Adapter】Supports the latest USB4 standard for up to 40Gbps of bandwidth, 8K video support, and up to 240W Fast charging; Support USB PD 3.1 Extended Power Range (EPR) including 140W, 180W and 240W. Your productivity is easily doubled. *Warm Tip: For the best fit, please make sure the USB-C port is fully exposed; thick or bulky cases may affect compatibility.
- 【Reliable Quality】Made from high-quality, scratch-resistant aluminum alloy, delicate metal frame. get excellent wear resistance and heat dissipation. Non-Slip Design for easier insertion and removal.
- 【Small and portable】Ultra thin size, ultra compact design, a petite physique that matches the minimalist design of your modern laptops. Coming with a slim figure that won't block the port next to the one it's plugged into.
- 【Compatibility】Widely work with most usb c, thunderbolt 4/3 devices. Compatibilit for ROG Ally, Steam Deck, Switch, MacBook Pro, Quest 3 VR headset, Surface Book 2, iPhone 18/17/16/15, Samsung Galaxy S23/S22/S21, Note 20/10, Google Pixel, ChromeBook, etc.
- 【Excellence Customer Support】If you experience any issues or cable failure, simply reach out to our dedicated customer service team. We provide a prompt response within 18 hours, providing you with the necessary technical support to resolve any concerns.
Which hardware is required for a mobile robot?
Start with the functions your existing base already provides, then fill the gaps. A practical mobile build normally includes the following.
Environment sensor
Choose a sensing modality that matches the map you need and the environment in which the robot will operate. A planar lidar can produce laser scans for a 2D map. An RGB-D camera supplies color and depth; a stereo camera derives depth from paired images. Multiple sensors can be combined, but every added stream increases calibration, bandwidth, timing, and compute work.
Continuous motion estimate
The mapper needs a smooth estimate between sensor updates. Wheel encoders are common on differential-drive bases, but Nav2 explicitly states that encoders are not required. IMU, camera, lidar, or other sources can provide odometry. If encoders and an IMU are both available, a fusion configuration can produce a smoother estimate than either source alone.
Onboard compute
The computer must run the robot middleware, sensor drivers, odometry, mapping or localization, and any navigation or perception nodes at the required rate. Size it from camera count, image resolution, lidar rate, algorithm, and concurrency rather than from a universal “SLAM computer” recommendation. CPU-only and integrated-graphics mobile-robot tracks are documented, while more demanding stereo and 3D pipelines use GPU-capable systems.
Persistent storage
Use non-volatile storage when maps or a mapping database must survive a power cycle. RTAB-Map’s documented default database path is ~/.ros/rtabmap.db. Storage capacity depends on whether you retain only an occupancy grid, a feature graph, raw images, depth, or other sensor data.
Power, data, mounting, and timing
Provide regulated power for every sensor and computer, interfaces that match their bandwidth, rigid mounts, and a known relationship between each sensor and the robot base. Ethernet, USB, and serial links have different cable and power limits. A good sensor with a loose mount or inconsistent timestamps can produce a worse map than a less expensive sensor that is rigidly mounted and correctly calibrated.
Do you need lidar for robot spatial memory?
No. Lidar is a viable sensing choice, not a universal requirement. Nav2’s official state-estimation documentation says: “There is no requirement on using a LIDAR on your robot to use the navigation system.” The important question is whether your selected mapper and localization system can consume the sensor data and whether the environment provides features that sensor can observe.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Fix the driver behind crashes, sound loss and screen glitches3Clear out junk files and repair common Windows errorsAlso popular on Amazon
- 5-in-1 Connectivity: Equipped with a 4K HDMI port, a 5 Gbps USB-C data port, two 5 Gbps USB-A ports, and a USB C 100W PD-IN port. Note: The USB C 100W PD-IN port supports only charging and does not support data transfer devices such as headphones or speakers.
- Powerful Pass-Through Charging: Supports up to 85W pass-through charging so you can power up your laptop while you use the hub. Note: Pass-through charging requires a charger (not included). Note: To achieve full power for iPad, we recommend using a 45W wall charger.
- Transfer Files in Seconds: Move files to and from your laptop at speeds of up to 5 Gbps via the USB-C and USB-A data ports. Note: The USB C 5Gbps Data port does not support video output.
- HD Display: Connect to the HDMI port to stream or mirror content to an external monitor in resolutions of up to 4K@30Hz. Note: The USB-C ports do not support video output.
- What You Get: Anker 332 USB-C Hub (5-in-1), welcome guide, our worry-free 18-month warranty, and friendly customer service.
When a 2D lidar build makes sense
For an indoor robot that needs a navigable floor-plan map, a 2D lidar, continuous odometry, onboard computer, and saved-map storage are a straightforward combination. Nav2 documents a workflow in which laser scans feed SLAM Toolbox during mapping and AMCL during localization on an existing map. This is an example architecture, not a rule that every 2D robot must use those packages.
Before selecting a lidar, verify the actual ROS 2 driver, scan field of view, mounting height, minimum and maximum range, physical clearance, update rate, and power connector. The available evidence does not establish one universally suitable lidar model or SKU.
Limits of a saved lidar map
A saved map describes previously observed structure; it does not tell the robot about a chair, person, or box that moved after mapping. Keep a live obstacle sensor in the navigation loop. The global map supports positioning and route planning, while current scans or depth observations support collision avoidance.
Can a depth camera replace lidar?
Often, yes, if your software supports RGB-D input and the scene supplies usable depth. RTAB-Map documents RGB-D mapping with synchronized color and depth images plus odometry. A concrete mobile-robot blueprint pairs a Clearpath Jackal configuration with an Intel RealSense D435i; that implementation should not be treated as a compatibility guarantee for every host, driver, or ROS distribution.
RGB-D hardware checklist
- Confirm the camera’s supported ROS 2 driver and message types.
- Provide enough USB bandwidth and stable power for the camera’s selected resolution and frame rate.
- Mount the camera rigidly, with a field of view that sees walls, furniture, and navigational obstacles rather than only the floor.
- Use the camera’s calibration data and verify that color, depth, and odometry timestamps are aligned.
- Plan for lighting and range limits. Active depth can degrade outdoors or on reflective, transparent, or very dark surfaces.
Depth cameras can reduce the need for a scanning lidar, but they do not remove the need for live obstacle observations, transforms, calibration, or a motion estimate.
Is stereo vision enough without lidar or active depth?
It can be, but passive stereo is more dependent on scene texture, lighting, motion, and camera stability. Nav2 describes a passive-stereo navigation path on a Jetson without lidar or active-depth sensors, while identifying glass, featureless corridors, thin obstacles, vibration, abrupt motion, and changing illumination as challenges.
RTAB-Map treats stereo as a distinct input path. Use a rigid stereo mount, accurate intrinsic and extrinsic calibration, and a baseline appropriate for the distances you need to measure. A stereo pair that sees only blank walls or suffers from vibration may provide too few reliable features for stable localization even when its cameras meet their nominal resolution.
Also popular on Amazon
- Sleek 7-in-1 USB-C Hub: Features an HDMI port, two USB-A 3.0 ports, and a USB-C data port, each providing 5Gbps transfer speeds. It also includes a USB-C PD input port for charging up to 100W and dual SD and TF card slots, all in a compact design.
- Flawless 4K@60Hz Video with HDMI: Delivers exceptional clarity and smoothness with its 4K@60Hz HDMI port, making it ideal for high-definition presentations and entertainment. (Note: Only the HDMI port supports video projection; the USB-C port is for data transfer only.)
- Double Up on Efficiency: The two USB-A 3.0 ports and a USB-C port support a fast 5Gbps data rate, significantly boosting your transfer speeds and improving productivity.
- Fast and Reliable 85W Charging: Offers high-capacity, speedy charging for laptops up to 85W, so you spend less time tethered to an outlet and more time being productive.
- What You Get: Anker USB-C Hub (7-in-1), welcome guide, 18-month warranty, and our friendly customer service.
What compute should you buy for visual or 3D mapping?
Choose compute after selecting the mapping and perception stack. Camera count, resolution, depth processing, loop-closure work, neural inference, and simultaneous navigation determine CPU, GPU, RAM, and storage demand.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Clear out junk files and repair common Windows errorsFree Scan →Moderate visual workloads
Documented mobile-robot tracks exist that use a CPU or integrated graphics. These can suit a single RGB-D or stereo stream when the selected algorithm and resolution fit their throughput. Leave headroom for navigation, logging, and safety nodes instead of allocating every CPU cycle to SLAM.
GPU-oriented workloads
Nav2 demonstrates a stereo path on an NVIDIA Jetson AGX Orin. NVIDIA’s current Isaac ROS mapping documentation specifies a 128+ GB NVMe SSD for its supported Jetson configuration. That capacity belongs to that documented stack and platform; it is not a minimum requirement for SLAM in general.
A published TargetBench research platform combines a Livox Mid-360 lidar, an OAK-D Pro W stereo camera, a Jetson AGX Orin, a dedicated battery, Ethernet for the lidar, and USB for the camera. It illustrates one combined 3D-and-vision design, not a universal shopping list or a guarantee that those parts work together on every software version.
Do you need wheel encoders, an IMU, or both?
You need a dependable continuous motion estimate, but no single odometry sensor is compulsory. Determine what the base, camera, or lidar driver already publishes before buying another sensor.
Free tools Windows power users keep installed
One-click scans. No signup required.
Wheel encoders
Encoders are a practical source for wheeled robots, especially differential-drive platforms. They provide high-rate relative motion but accumulate error when wheels slip, traverse uneven surfaces, or have mismatched diameters. If the base already exposes calibrated odometry, adding separate encoders may duplicate hardware without improving the estimate.
IMU
An IMU contributes angular velocity and acceleration, which can help during turns and short periods when visual or wheel measurements are weak. It does not by itself provide a globally anchored map; bias and drift must be handled by the estimator.
Also popular on Amazon
- Portable and powerful USB-C HUB: BENFEI USB Type-C HUB, with super-soft and knot-free silicone woven design cable, meets most mobile office needs. Compact, lightweight, stylish, and powerful portable USB C Hub equipped with 1 x HDMI port, 1 x 100W charging, and 3 x USB ports. 18-month warranty, 24-hour response, to ensure you feel at ease when using our product.
- Design centered on comfort and reliability: Thanks to BENFEI's end-to-end in-house cable production capability, in-house PCBA and assembly capability, using the industry's most advanced silicone woven design and process, 20cm cable in length, no knots, super-soft, the HUB is easy to use in all scenarios: laptop, tablet, stand etc. Super-soft, 25000+ life cycles, to meet your daily carrying and office needs.
- 100W Charging: Support up to 90W USB C pass-through charging via Type-C port to keep your laptop powered. 10W is reserved for other interface operations. No data and video function on the Type-C port.
- 4K HDMI Display: The HDMI port supports media display at resolutions up to 4K 30Hz, keeping every incredible moment detailed and ultra vivid. Please note that the C port of the Host device needs to support video output.
- Transfer Files in Seconds: Transfer files and from your laptop at speeds up to 10 Gbps with USB A 3.2 port. Extra 2 USB A 2.0 ports are perfectly for your keyboards and mouse.
Sensor fusion
When encoder and IMU data are both available, Nav2’s setup guidance describes fusing them for a smoother estimate. The fusion node must know each sensor’s frame, covariance, timestamp behavior, and mounting orientation. Incorrect covariances or a reversed axis can make a fused result less reliable than a single correctly configured source.
How important are coordinate frames and timestamps?
They are as important as the sensor choice. A typical navigation chain is map → odom → base_link → sensor. The global localization component updates map → odom; local odometry maintains the continuous odom → base_link transform; a static transform places each lidar or camera relative to base_link.
Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minutePC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Check that every message names the intended frame, that transforms exist at the message timestamp, and that clocks remain synchronized across computers. A map can appear to warp when the real problem is a delayed camera frame, an incorrect sensor offset, or a transform published in the wrong direction.
For a specific cuVGL camera system, NVIDIA publishes a target image rate of at least 5 Hz, synchronization within a stereo camera of ±10 microseconds, and synchronization across stereo cameras of ±100 microseconds. Those are requirements for that NVIDIA stack in 2026, not universal visual-SLAM thresholds. Use the requirements published for your own software and camera combination.
Where is the map stored, and what survives a restart?
Separate three kinds of memory:
- Persistent disk storage holds a database or exported map after shutdown. RTAB-Map documents
~/.ros/rtabmap.dbas its default database location and can load a saved occupancy grid at startup. - Active working memory contains the nodes currently available for fast mapping and localization. RTAB-Map can bound this memory and move older nodes into its long-term database, retrieving them when the robot returns to a region.
- Semantic data records labels, landmarks, or application-specific information such as “charging station” or “conference room.” RTAB-Map accepts user data and landmarks and can label goal nodes.
Bounding active memory has a trade-off: older regions may disappear from the occupancy grid that Nav2 currently sees even though RTAB-Map retains them in its graph. Ensure your navigation design can request or rebuild the needed region before relying on a very large environment.
Which build path fits your goal?
| Goal | Typical sensing | Motion estimate | Software pattern | Main risks to check |
|---|---|---|---|---|
| 2D indoor navigation map | 2D lidar laser scan | Wheel, IMU, visual, or lidar odometry | SLAM Toolbox for mapping; AMCL or another localization system on a saved map | Driver support, scan coverage, wheel slip, live obstacle sensing |
| RGB-D geometry and repeat localization | RGB-D camera | Camera or base odometry, optionally fused with IMU | RTAB-Map RGB-D pipeline | Depth range, lighting, USB bandwidth, synchronization |
| Passive-stereo navigation | Calibrated stereo pair | Visual, wheel, IMU, or fused odometry | RTAB-Map stereo or a Nav2 stereo configuration | Glass, blank walls, thin obstacles, vibration, changing light |
| Higher-demand 3D or multimodal mapping | 3D lidar plus stereo or RGB-D | Fused wheel, IMU, visual, and lidar sources as supported | GPU-capable mapping and perception stack | Compute heat and power, bandwidth, calibration, storage growth |
How do you choose parts without overbuying?
- Define the memory outcome. Decide whether you need a 2D occupancy grid, 3D geometry, repeat localization, named places, or object and room semantics.
- Inventory existing outputs. Record the base’s odometry topic, frame names, timestamps, available power, and computer interfaces before purchasing encoders or an IMU.
- Select the sensor modality. Compare lidar, RGB-D, and stereo against lighting, texture, glass, obstacle height, field of view, and mounting constraints.
- Choose the software first. Confirm that the mapper, localization system, ROS 2 distribution, and sensor drivers support the exact message formats and rates you plan to run.
- Budget compute and storage. Include navigation, logging, and safety processing. Treat stack-specific figures such as NVIDIA’s 128+ GB NVMe requirement as scoped specifications, not generic minimums.
- Design integration details. Reserve regulated power, data bandwidth, rigid mounts, calibration access, and synchronized clocks. Publish and inspect the complete frame tree.
- Plan persistence and recovery. Decide where databases and exported maps live, how they are backed up, and how the robot handles a missing or partially corrupted map after restart.
What mistakes cause unreliable spatial memory?
- Buying a sensor before checking its driver: A physically compatible device may lack support for your ROS 2 distribution or required message type.
- Treating a map as a live obstacle detector: Static map data cannot describe furniture or people that moved after mapping.
- Ignoring transforms: A few centimeters of wrong sensor extrinsics can create doubled walls or drifting scans.
- Using unsynchronized clocks: Timestamp skew makes otherwise good observations appear to come from different poses.
- Underestimating power and thermals: A computer that throttles or resets under camera and GPU load cannot maintain localization.
- Confusing geometry with semantics: A map does not automatically know that a region is a kitchen or that a shape is a particular object. Add labels, landmarks, or a perception system for those tasks.
- Over-bounding working memory: Removing old nodes from the active occupancy grid can leave Nav2 without the region it needs, even when the long-term database still contains it.
How should you compare candidate hardware?
Use six comparison axes rather than a single “best sensor” ranking:
- Required output: 2D map, 3D geometry, localization, or semantic memory.
- Environment: lighting, texture, glass, reflective surfaces, dust, and moving obstacles.
- Coverage: field of view, scan plane or depth range, blind spots, and obstacle height.
- Motion and timing: available odometry, calibration, frame rates, and clock synchronization.
- Compute and storage: workload-specific CPU/GPU demand, RAM, thermal envelope, and database growth.
- Integration: mount, power, cabling, ROS driver, middleware, and version support.
How we picked: These criteria follow the documented roles of sensing, odometry, frame transforms, synchronization, mapping, navigation, and persistence. They distinguish universal functions from examples tied to a particular platform or software stack.
Also popular on Amazon
- 【Read before You Buy】 Note: This adapter is USB 2.0 version and cannot transmit HDM, VGA, or other video signals. Nevertheless, they are fully capable of charging devices,duplicating files, and can be seamlessly integrated with USB-C earphones, wireless network adapters, and SD card adapters.
- 【Easy to Use and Safe】Thanks to modern manufacturing techniques, we've made these adapters sufficiently compact so that they can remain inserted in your device without noticeable protrusion or inconvenience. This means you don't have to constantly plug and unplug them, reducing the likelihood of wear and misalignment, and making them more convenient to carry around.
- 【Reliable Quality】Engineered for enhanced durability, durable aluminum alloy shell material,get excellent wear resistance and heat dissipation,the male USB C and female USB connectors & the male USB and female USB C connectors are designed to be plugged and unplugged up to 20000 times. Provides data transfer, device charging, memory space expansion, and additional computer peripherals.
- 【Think more for you】Do you need to connect a USB-C-enabled device to an odd device that only has USB-A ports? Then these USB-A male to USB-C female adapters are made for you.Simply insert the adapter into the USB-A port on the device, and you can instantly connect your USB-C device to the device that originally didn't have the correct connection port.
- 【Trusted xiwxi Quality and Service】As a responsible brand seller, you will receive xiwxi's 12-month Quality assurance and lifetime friendly customer service, you're not just buying a product,you're investing in peace of mind with all-round quality and exceptional customer service.
Can spatial memory include room names and object identities?
Not from a geometric map alone. A lidar or depth map can preserve shape and location, but room names and object identities require application-level labels or a perception model. RTAB-Map supports user data, landmarks, and labels attached to map nodes, which provides a place to store that semantic layer alongside metric geometry.
A useful design keeps these layers separate: the navigation map answers “where can the robot drive?”, localization answers “where am I?”, and semantic records answer “what is this place or object called?”
What evidence exists for long-term indoor mapping?
A 2017 journal study by Mathieu Labbé and François Michaud reported 10.5 km of indoor travel over 11 sessions while encountering 139 people. That figure belongs to the conditions and system described in that study; it is not a performance promise for every robot assembled from the parts above.
Recommended Free Tools
Frequently Asked Questions
Can a robot use a saved map if its mapping computer is turned off?
Yes, if the map has been exported to storage and a separate onboard computer can run localization and navigation. The robot still needs live sensor observations and odometry to estimate its current pose and detect obstacles.
How much storage does a robot map need?
There is no universal capacity. A compact occupancy grid uses far less space than a database retaining images, depth frames, feature descriptors, and repeated sessions. Measure the data products your selected stack saves and keep free space for logs and recovery copies.
Will transparent glass appear correctly in a lidar or camera map?
Not reliably. Glass can produce weak, missing, or misleading returns for lidar and stereo, while depth cameras can also struggle with transparent or reflective surfaces. Add a sensor and navigation rule that covers the blind spot rather than assuming one modality sees every obstacle.
What should happen when the robot cannot localize on its old map?
Stop or switch to a safe recovery behavior, then collect fresh observations while controlling motion. Check the frame tree, timestamps, sensor exposure, and whether the environment changed; relocalization is a software and operating-procedure problem as well as a hardware one.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
