Simultaneous localization and mapping (SLAM)

Overview

The sixteen cards use SLAM, or the closely related task of building a map while moving, in different ways. LVI-SAM fuses a lidar, a monocular camera and an IMU in one factor graph, with loop closure 2104.10831 §2.1 2104.10831 §2.2.4. The racing survey lists SLAM as one way to localise a car, next to Monte Carlo (AMCL) localisation 2202.07008 §2.1. The benchmark card evaluates SLAM systems directly. GS3LAM and SplaTAM are full RGB-D SLAM methods: each estimates camera poses and a map at the same time. GS3LAM also builds a semantic map 2603.27781 §3.4 2312.02126 §3. NeRF-SLAM is monocular: a dense SLAM tracker supplies poses and depth, and a radiance field is fitted to them in parallel 2210.13641 §3.3. NeuSE is an object-level SLAM pipeline: it adds object-level pose constraints to a pose graph, and the graph's inter-frame edges can come from NeuSE or from another source 2303.07308 §4.1. iSDF maps with poses supplied by an external tracker, and says it does not do full SLAM 2204.02296 §4. The navigation paper does not evaluate SLAM at all, but its evaluation rules apply to any agent that explores a map before it is tested 1807.06757 §3. Replica is a simulated indoor dataset, not a method. Its own capture pipeline used an in-house SLAM system to estimate camera poses 1906.05797 §3. SuMa++ is a lidar-only surfel SLAM system with semantic labels from a neural segmenter. It uses the labels to find moving objects in the map and to weight the scan alignment 2105.11320 §III-E 2105.11320 §III-F. Kimera runs visual-inertial odometry online and builds a labelled mesh as it goes. It parses a five-layer 3D scene graph after the run 2101.06894 §2 2101.06894 §3. R3LIVE builds its map geometry from a lidar-inertial step, and colours that map with a camera step 2109.07982 §II. ConceptFusion is a mapping method rather than a SLAM one: it takes posed RGB-D frames and fuses per-pixel features into a 3D map, which it can then query by text, image, click or audio 2302.07241 §3.1 2302.07241 §3.3.

The benchmark card describes the gap it fills. Most public SLAM datasets are built for driving or aerial robots, and the few for ground robots lack either a rich sensor set or hard scenes such as lifts and darkness 2112.13659 §1 2112.13659 §2.2.

For the SLAM systems, the shared invariant is how trajectories are judged. Each estimated trajectory is scored against a reference trajectory from motion capture, a laser tracker or an RTK/INS system 2112.13659 §4 2112.13659 tab. 3. LVI-SAM scores against a GPS receiver, as RMSE, and also reports end-to-end translation and rotation errors 2104.10831 §3 2104.10831 tab. 2. GS3LAM also reports ATE RMSE, on Replica and on ScanNet 2603.27781 tab. 3 2603.27781 app. C.3. SplaTAM reports ATE RMSE on Replica, TUM-RGBD, ScanNet and ScanNet++ 2312.02126 tab. 1. NeuSE reports ATE RMSE against a reference from Vicon or the simulator 2303.07308 §7.1 2303.07308 §7.2. iSDF scores its map instead, as SDF error against a ground-truth grid built from the mesh 2204.02296 §5.1 2204.02296 §5.3. The navigation paper uses a different score, based on whether the agent stops at the goal and on its path length 1807.06757 §4, so the families of numbers measure different things. R3LIVE measures drift at the start point of each loop, against an ArUco board, and relative pose error against RTK 2109.07982 §VI-C 2109.07982 tab. II 2109.07982 tab. III. Kimera scores EuRoC trajectories with ground truth after SE(3) alignment, and scores its mesh by RMSE 2101.06894 §4.2 2101.06894 tab. 3. ConceptFusion reports 3D mIoU for query results, not a trajectory 2302.07241 tab. 1.

Coverage is still thin. M2DGR and LVI-SAM each compare several visual, LiDAR or fused systems on their own recorded sequences 2112.13659 §4 2104.10831 §3. The RGB-D methods compare against several dense and feature-based baselines, but on simulated and handheld data, and some baseline numbers are taken from other papers: SplaTAM takes its Replica, TUM-RGBD and ScanNet baselines from Point-SLAM 2312.02126 §4. NeuSE and iSDF report their own systems against a few baselines. The Gaussian-splatting survey repeats SLAM numbers from the original papers, under settings that differ between them 2401.03890 tab. S4 2401.03890 §6. The racing survey, the small-scale car survey and the navigation paper cover SLAM-related work without running any SLAM system. Kimera's real-world evidence is qualitative, and its numbers come from simulated humans and the EuRoC sequences 2101.06894 §4.1 2101.06894 §4.10. R3LIVE's results come from the authors' own rig, against baselines the authors ran 2109.07982 §VI-A 2109.07982 §VI-D. ConceptFusion's robot tests are qualitative, with no success rates 2302.07241 §4.3. NeRF-SLAM compares against iMAP, Nice-SLAM and two fusion baselines, all on simulated Replica scenes, and reports depth error and PSNR on rendered views rather than trajectory error 2210.13641 §4.2 2210.13641 tab. 1.

Variants

Visual SLAM

ORB-SLAM3 in monocular mode without an IMU, with pinhole, fish-eye and thermal cameras. Also CubemapSLAM and VINS-Mono. Multicol-SLAM is named in the text but has no numbers 2112.13659 §4 2112.13659 tab. 5. Card: M2DGR.

LiDAR SLAM

A-LOAM, LeGO-LOAM, LINS and LIO-SAM 2112.13659 §4. Card: M2DGR. For the LiDAR-inertial case with a card, see lidar-inertial-odometry.

Lidar-visual-inertial SLAM

LVI-SAM runs a visual-inertial sub-system and a lidar-inertial sub-system. Each gives the other a starting point and can flag its own failure. Both feed one iSAM2 factor graph with IMU, visual-odometry, lidar-odometry and loop-closure constraints 2104.10831 §2.1 2104.10831 §2.3. Loop candidates come from DBoW2 on the camera side and are checked by scan matching on the lidar side 2104.10831 §2.2.4 2104.10831 §2.3. It needs a 3D lidar, one monocular camera and one IMU, with known extrinsics; the paper does not describe how to calibrate them 2104.10831 §2.1. Author-reported, one run per method: on Jackal, RMSE against GPS of 0.67 m with loop closure; on Handheld, where the lidar baselines fail, 0.83 m with loop closure and 7.87 m without 2104.10831 tab. 2 2104.10831 §3.2 2104.10831 §3.3. In its Urban ablation, lidar depth on visual features lowers translation error from 45.42 m to 32.18 m without loop closure, and raises rotation error from 6.41° to 7.66° 2104.10831 tab. 1. Card: LVI-SAM.

Colourised lidar-inertial-visual mapping

R3LIVE splits the work in two. A lidar-inertial step runs an error-state iterated Kalman filter on point-to-plane residuals, and builds the geometry of one global map 2109.07982 §II 2109.07982 §IV. A visual-inertial step then tracks map points with Lucas-Kanade optical flow. It refines the state with a reprojection residual, and then with a photometric residual against the stored colours 2109.07982 §V-A 2109.07982 §V-B1. Each map point keeps a colour and its variance, and each observation is fused into them by a Bayesian update 2109.07982 §V-C. The one texture-poor test is a narrow passage with bare white walls, where drift at the end is 1.62 degrees and 4.57 cm 2109.07982 §VI-B. It needs a 3D lidar, a global-shutter camera and an IMU. The state also estimates the camera intrinsics, the camera-IMU extrinsic and their time offset online 2109.07982 §III-A 2109.07982 §VI-A. The paper does not say whether the visual step can run without the lidar map, which supplies its depth 2109.07982 §II 2109.07982 §V. Author-reported: on four campus trajectories of 1191 to 1524 m, drift at the start point is 0.093 to 0.164 m in translation and 0.285 to 3.925 degrees in rotation. Each loop closes without a loop-closure step 2109.07982 tab. II 2109.07982 §VI-C. On a seaport sequence with pedestrians, the 50 m relative pose error is 0.99 degrees and 2.25% for R3LIVE-HiRes, against 2.11 degrees and 13.73% for LVI-SAM, which the authors ran with a modified lidar front end 2109.07982 tab. III 2109.07982 §VI-D. The visual step takes 13.63 ms per frame on a desktop and 33.38 ms on the onboard computer of a drone, at 1280 x 1024 and 0.10 m point spacing. At 0.01 m spacing the onboard time is 166.08 ms 2109.07982 tab. IV. Meshing is offline and not real-time 2109.07982 §VII-A. The released code is GPL-2.0; see the card's Paper vs code section before reusing anything from it. Card: R3LIVE.

Metric-semantic mesh and scene-graph SLAM

Kimera runs visual-inertial odometry online, with fixed-lag smoothing in GTSAM. It builds a semantic mesh as it goes 2101.06894 §3.1 2101.06894 §3.2 2101.06894 §3.3. Loop closures pass a pairwise-consistency test, and the optimised deformation graph then corrects the mesh 2101.06894 §3.4. Objects, places, rooms and the building are parsed from the mesh after the run. Incremental parsing is listed as future work 2101.06894 §3 2101.06894 §8. Humans are tracked with pose graphs built from GraphCMR estimates 2101.06894 §3.5. It needs a stereo or RGB-D camera, an IMU, and 2-D semantic labels per keyframe. On real data the labels come from Mask R-CNN 2101.06894 §3 2101.06894 §4.1. Author-reported: on EuRoC V1_1, Kimera-VIO's ATE is 0.05 m, against 0.09 m for OKVIS and 0.34 m for MSCKF. With loop closure, Kimera-PGMO gives 0.05 m on V1_1 and 0.09 m on MH_1 2101.06894 tab. 1. Mesh RMSE with 60 simulated humans and ground-truth poses is 0.19 m without dynamic masking and 0.06 m with it 2101.06894 tab. 5. Hierarchical A* planning in a scaled office, with one to six buildings, takes 0.340 s to 1.244 s, against 420 s to 21,389 s with A* on the ESDF 2101.06894 §6.3 2101.06894 tab. 9. Keyframe processing takes 51 ms, the per-frame mesh under 7 ms, and building the Subway scene graph about 30 minutes 2101.06894 §4.8. The paper gives no CPU model for those timings 2101.06894 §4.8. The EuRoC trajectories are aligned with SE(3), not Sim(3), which the paper says would lower the error 2101.06894 §4.2. Its two Kimera-Semantics voxel sizes differ, 0.05 m in §3.6 and 0.1 m for the TX2 setup 2101.06894 §3.6 2101.06894 §4.9. The repository at the linked commit holds only documentation, so no module code has been checked; see the card's Paper vs code section. Card: Kimera.

Semantic surfel LiDAR SLAM

SuMa++ keeps the surfel map of SuMa and gives each surfel a semantic label and a stability value 2105.11320 §III-B 2105.11320 §III-C. A segmenter labels each lidar point, and the label image is cleaned before use 2105.11320 §III-C 2105.11320 §III-D alg. 1. A surfel whose label conflicts with a movable class, such as a car or a person, loses stability. The paper removes such surfels after several observations 2105.11320 §III-E eq. 1. Vegetation and other static classes are not penalised 2105.11320 §IV-C. The same labels down-weight residuals in the scan alignment 2105.11320 §III-F eq. 6. Author-reported: average relative translation error on KITTI odometry training sequences 00–10 falls from 0.83% for SuMa to 0.70% for SuMa++ 2105.11320 tab. II. On road sequence 35, a highway with many moving cars, it falls from 26.8% to 1.11% 2105.11320 tab. I. Deleting every movable-class surfel (SuMa_nomovable) is also a baseline: it reaches 1.22% on sequence 35 but 9.24% on the odometry training average 2105.11320 tab. I 2105.11320 tab. II. The segmenter is trained on sequences 00–10 except 08, so Table II is in-sample; the only held-out number is the server test result, 1.06% against SuMa's 1.39% 2105.11320 §IV 2105.11320 §IV-B. It needs a labelled lidar training set and a trained segmenter. Reported timings are about 75 ms for the segmenter and about 48 ms for surfel mapping per scan, on one Quadro P4000 2105.11320 §IV. Card: SuMa++.

Monocular dense SLAM with a radiance field

NeRF-SLAM runs two threads in parallel. The tracker is DROID-SLAM, which gives keyframe poses, dense depth maps and their marginal covariances 2210.13641 §3.1 2210.13641 §3.3. The mapper fits an Instant-NGP radiance field to the keyframes it receives 2210.13641 §3.2 2210.13641 §3.4. The depth term of its loss is weighted by the depth covariance, so uncertain depths count for less 2210.13641 eq. 4 2210.13641 §3.2. The poses are refined jointly with the field 2210.13641 eq. 3. The mapper uses every keyframe it has received, with no sliding window 2210.13641 §3.3. It needs only a monocular camera, with no depth sensor. Author-reported on Replica, its PSNR is the highest in all eight scenes, and its average depth error, 4.49 cm, is above the 4.08 cm of Nice-SLAM with ground-truth depth 2210.13641 tab. 1. The headline gains of up to 179% in PSNR and 86% in depth error are against the Nice-SLAM row without depth 2210.13641 §4.3 2210.13641 tab. 1. In an ablation on a synthetic scene, raw monocular depth lowers depth error but makes PSNR about 3 dB worse than using no depth; weighting by covariance avoids that 2210.13641 §4.4. Speed is about 10 fps on 640 × 480 images on one RTX 2080 Ti with 11 GB of memory 2210.13641 §4.5 2210.13641 §3.4. The authors name dense correlation volumes and hierarchical grids as the memory cost 2210.13641 §5. Card: NeRF-SLAM.

RGB-D Gaussian-splatting SLAM

Both methods here keep the scene as 3D Gaussians and track each new frame against that map. They differ in what each Gaussian holds and in what input they need.

GS3LAM represents the scene as a set of 3D Gaussians, each with a semantic feature that a learned layer decodes to class scores 2603.27781 §3.2. Tracking optimises only the camera pose, with the field frozen. Mapping then refines the field on the current frame and on keyframes drawn at random 2603.27781 §3.4 2603.27781 §3.3.3. It needs one semantic label map per frame, from the dataset or from a segmentation model 2603.27781 §3.3.4. Author-reported: rendering PSNR 36.26 over eight Replica scenes 2603.27781 tab. 2, and tracking ATE 0.37 cm on Replica against SplaTAM's 0.36 cm 2603.27781 tab. 3. Its released rasterizer carries a non-commercial licence; see the card's Paper vs code section. Card: GS3LAM.

SplaTAM keeps a simpler map: each Gaussian has a colour, a centre, a radius and an opacity, with no semantic feature and a view-independent colour 2312.02126 §3 Gaussian Map Representation. For each frame it fits the camera pose with the map held fixed, adds Gaussians where the map is sparse or misses geometry, then refines all Gaussians over a window of keyframes 2312.02126 §3 SLAM System. It assumes one unposed RGB-D camera, with known intrinsics and dense depth 2312.02126 §1 [arxiv:2312.02126 §5 Limitations]. Author-reported: ATE RMSE of 0.36 cm on Replica, 5.48 cm on TUM-RGBD, 11.88 cm on ScanNet and 1.2 cm on ScanNet++, where ORB-SLAM3 scores 158.2 cm and Point-SLAM 343.8 cm 2312.02126 tab. 1. Its training-view rendering on Replica is below Point-SLAM's, at 34.11 dB against 35.17 dB 2312.02126 tab. 2. Per frame on Replica R0 on one RTX 3080 Ti, tracking takes 1.00 s and mapping 1.44 s 2312.02126 tab. 6. Card: SplaTAM.

A survey of 3D Gaussian splatting repeats SLAM numbers for these methods rather than running them. On Replica, its table gives SplaTAM 0.36 cm ATE against Point-SLAM 0.52 cm and NICE-SLAM 1.06 cm, and gives Gaussian-SLAM 3.27 cm 2401.03890 tab. S4. The survey says it compiles results from the original works, which use different settings 2401.03890 §6. Card: A Survey on 3D Gaussian Splatting.

Neural SDF mapping with given poses

iSDF trains a small MLP online on posed depth images, so that it approximates a signed distance field. It uses the nearest measured surface point in each batch as the supervision bound 2204.02296 §4.3. The paper assumes an external tracker supplies the camera poses 2204.02296 §4. Author-reported: SDF error below 6 cm at the end of each of six main sequences [arxiv:2204.02296 §6.2] 2204.02296 fig. 8, and about 33 ms per training iteration on an i7-8700K with an RTX 2080 2204.02296 §5.2 2204.02296 §6.3. Card: iSDF.

Object-level SLAM with learned codes

NeuSE gives each object a latent code, learned from a partial point cloud. The code is SE(3)- and scale-equivariant, so the relative pose of two views of an object is read from their codes with a closed-form Horn fit, not by rendering and optimising 2303.07308 §1 2303.07308 §3.2. The codes drive four things: data association, pose constraints in a pose graph, a sliding window of ten keyframes, and change detection in the object map 2303.07308 §4.1 2303.07308 §4.2 2303.07308 §4.3 2303.07308 §8. In the real experiments the baseline and base system is ORB-SLAM3 2303.07308 §7.2. Each object category needs its own trained model 2303.07308 app. A1, and the models are trained on simulated data only 2303.07308 §5.1. Card: NeuSE.

Open-set feature mapping

ConceptFusion stores a language-aligned feature vector in each map point. It fuses these with the same confidence-weighted average used for depth and colour. The confidence falls off with distance from the camera 2302.07241 §3.1 2302.07241 eq. 1 2302.07241 eq. 2. Each pixel's feature is built without training. A CLIP-style embedding of the whole frame is mixed with embeddings of the boxes around class-agnostic masks, and the mix is weighted by similarity and by a uniqueness term 2302.07241 §3.2 2302.07241 eq. 3 2302.07241 eq. 6. A query, given as text, an image, a clicked point or audio, is scored against each map point by cosine similarity 2302.07241 §3.3. It needs posed RGB-D frames and a mask proposer: Mask2Former with 100 proposals per image, or SAM 2302.07241 §3 2302.07241 app. A4. Feature extraction is the slow step, at 10 to 15 s per image on an RTX 3090, and the authors give 15 Hz for odometry and mapping 2302.07241 §3. Outdoors the paper uses LeGO-LOAM for odometry 2302.07241 app. A3. Author-reported, on UnCoCo, a tabletop set with 78 household objects. With structured text queries the 3D mIoU is 0.446, against 0.289 for supervised OpenSeg-3D, 0.128 for supervised LSeg-3D and 0.091 for zero-shot MaskCLIP-3D 2302.07241 tab. 1. With long free-form text it is 0.378, against 0.153 for OpenSeg-3D 2302.07241 tab. 2. With image queries it is 0.331, against 0.134 for LSeg-3D 2302.07241 tab. 3. Audio queries reach 64.29% accuracy on one set, against 7.14% for chance and 23.81% for AudioCLIP given privileged boxes 2302.07241 tab. 4. In 2-D it trails supervised LSeg on ScanNet, with mAcc 0.63 against 0.70 2302.07241 tab. 5. Removing the uniqueness term lowers ScanNet mAcc to 0.55 2302.07241 tab. 6. On Replica, SAM masks give 31.53 mAcc, against 24.16 for Mask2Former 2302.07241 tab. 8. Spatial queries such as distance go through an LLM parser, and the paper reports 84% for distance queries on 100 spatial queries 2302.07241 tab. 7 2302.07241 app. A8. Caveats. The abstract's "more than 40% margin" is not defined, and the table gives a 0.157 mIoU gap 2302.07241 §1 2302.07241 tab. 1. The pixel features are centred on foreground objects, and the paper says they do not handle negation 2302.07241 §6. The released extraction script uses SAM and drops the uniqueness term, and the LLM parser is not in the repository; see the card's Paper vs code section. Card: ConceptFusion.

Datasets for SLAM and mapping

Two cards describe data rather than a method. M2DGR is a ground-robot SLAM dataset with trajectory references, described under the visual and LiDAR variants above. Replica is a set of 18 indoor scene reconstructions with dense meshes, HDR textures and semantic labels, built with an RGB-D rig and an in-house SLAM system 1906.05797 §3 1906.05797 §4. The paper reports no SLAM benchmark on it 1906.05797 §Results. Its repository licence is for non-commercial research and education only, which is stricter than the paper's CC BY 4.0 licence 1906.05797 §Pitfalls. Card: Replica.

GNSS reference

RTKLIB is evaluated as a GNSS baseline, not as SLAM 2112.13659 tab. 5.

Localisation in racing

The survey names localisation methods for racing cars, with errors it reports from other papers: AMCL on LiDAR, NDT, extended Kalman or H-infinity filters that fuse odometry, IMU and GPS, and graph-SLAM with cone detection on Formula Student cars. It also names Google Cartographer and GMapping, adapted to large outdoor racetracks 2202.07008 §2.1. Card: Autonomous Vehicles on the Edge.

Localisation on small-scale research cars

This is a survey, not a SLAM method. It lists localisation and mapping as one perception task for small robot cars, and compares the methods used for it in a table of check and cross marks, with no measured values 2404.06229 §4.1.1 2404.06229 tab. IV. Central localisation, such as motion capture or markers, is described as expensive and hard to move, which matters for outdoor use 2404.06229 §3.5 2404.06229 §4.3.5. The survey's own baseline pipeline calls SLAM to update the map, but the paper does not test that pipeline 2404.06229 §4.5. Card: Autonomous Driving Small-Scale Cars.

Map and exploration in navigation evaluation

This is not a SLAM method. The navigation paper asks that the prior exploration of the test environment be stated, because it changes what a score measures. It names three regimes: no prior exploration; pre-recorded exploration supplied with the benchmark; and exploration by the agent under a time budget. In the last regime, the paper suggests reporting performance as a function of the budget, and gives 500 m, 1,000 m and 2,000 m as examples 1807.06757 §3. It also asks that any perfect odometry or GPS assumption be stated 1807.06757 §2, and it asks for research on the internal representation an agent builds while it moves, without recommending a form for it 1807.06757 §7. Card: On Evaluation of Embodied Navigation Agents.

When to use which

The first rows follow the M2DGR scenes, then the method rows, then the navigation row. Results are author-reported, per sequence or per scene, and not comparable across platforms.

Implementation sketch

The benchmark cards do not give the internal steps of the SLAM systems they evaluate. For GS3LAM's per-frame loop, read its card. For SplaTAM's per-frame loop, read its card under Method. For NeuSE's data association order, read its card under Method. For LVI-SAM's update cycle, read its card under Method. For R3LIVE's per-scan and per-image steps, Kimera's online pipeline, ConceptFusion's per-pixel features and NeRF-SLAM's two threads, read their cards under Method. This sketch shows only the evaluation core that the benchmark uses to score one run.

# ATE evaluation of one run  [arxiv:2112.13659 §4]
# Input:  estimated poses and reference poses, with timestamps
# Output: one ATE value in metres for the sequence

def evaluate_run(estimated, reference):
    # Pair poses that share a timestamp; the paper does not say how  [§4]
    pairs = match_by_timestamp(estimated, reference)
    # Align the estimate to the reference before scoring  [§4]
    aligned = align_with_evo(pairs.estimated, pairs.reference)
    # Combine the per-pose errors into one number; not spelled out  [§4]
    return ate(aligned, pairs.reference)

Contradictions

• LiDAR versus vision. The authors say LiDAR-based methods generally beat vision-based methods 2112.13659 §4. On Hall05, VINS-Mono gives 0.646 m, which is below every LiDAR value there (1.010–1.065 m) 2112.13659 tab. 5.
• LVI-SAM across two papers. R3LIVE reports LVI-SAM at 2.11 degrees and 13.73% at 50 m on seaport sequence (a), with a modified lidar front end 2109.07982 tab. III. The LVI-SAM paper reports RMSE against GPS of 0.67 m and 0.83 m on two of its own datasets, with loop closure 2104.10831 tab. 2. The metric, data and LVI-SAM version differ, so the two sets of numbers cannot be compared, and this page does not reconcile them.
• Tracking accuracy in GS3LAM. The paper calls its tracking competitive, but its own tables do not show it as the best. On Replica, SplaTAM's ATE is 0.36 cm against GS3LAM's 0.37 cm 2603.27781 tab. 3. On ScanNet, NICE-SLAM's ATE is 10.70 cm against GS3LAM's 11.24 cm 2603.27781 app. C.3.
• Replica versus ScanNet in SplaTAM. SplaTAM's ATE on Replica (0.36 cm) is the lowest in its table, against DROID-SLAM at 0.38 cm. On ScanNet it is higher than NICE-SLAM, at 11.88 cm against 10.70 cm 2312.02126 tab. 1. GS3LAM's appendix gives the same two ScanNet values 2603.27781 app. C.3.
• NeuSE improvement figure. The text gives a 65% ATE improvement on the lower half of one trajectory 2303.07308 §7.1. Table 3 gives reductions against the point-cloud baseline for the full trajectory of about 58% (planar) and 66% (non-planar) 2303.07308 tab. 3. The two figures measure different things, so cite the table values.
• Speed claims. GS3LAM's 109 FPS is a rendering rate 2603.27781 app. C.2. iSDF's 33 ms is one training iteration, and the paper does not evaluate frame rate or tracking 2204.02296 §5.2 2204.02296 §6.3. NeuSE's about 4 fps is a full localization run in a crowded setting, with sequential processing and no software optimisation 2303.07308 §7. None of these is a comparable SLAM loop rate, and they should not be compared as if they were.
• LVI-SAM degeneracy rule. The paper says the lidar-odometry factor is not added when a scan match is degenerate 2104.10831 §2.3.2. The released code keeps the factor and projects out the degenerate directions; see the LVI-SAM card's Paper vs code section.
• LVI-SAM depth ablation. The text says lidar depth reduces error 2104.10831 §3.1.1. In the same ablation, rotation error rises from 6.41° to 7.66° 2104.10831 tab. 1. Both are reported as they stand.
• SplaTAM improvement over Point-SLAM. The 3DGS survey's text says SplaTAM cuts error by "~50%" from 0.52 cm to 0.36 cm 2401.03890 §6.1. Its own table gives those values, and (0.52 − 0.36) / 0.52 is about 31% 2401.03890 tab. S4. Cite the table value.
• NeRF-SLAM frame rate. The paper gives about 12 fps in one sentence and 10 fps overall, with tracking at about 15 fps and mapping at about 10 fps 2210.13641 §4.5. Quote the lower figure unless you measure your own setup.
• Different metrics. The survey reports lateral and longitudinal errors and RMSE for racing cars 2202.07008 §2.1. M2DGR, GS3LAM, SplaTAM and NeuSE report ATE for trajectories 2112.13659 §4 2603.27781 tab. 3 2312.02126 tab. 1 2303.07308 tab. 5. NeRF-SLAM reports depth L1 and PSNR on rendered views 2210.13641 tab. 1. LVI-SAM reports RMSE against GPS 2104.10831 tab. 2. SuMa++ reports KITTI relative translation error as a percentage of distance travelled 2105.11320 tab. I. iSDF reports SDF error and collision-cost error for a map 2204.02296 §5.3. The navigation paper scores success and path length, not ATE 1807.06757 §4. R3LIVE reports drift and relative pose error, Kimera reports ATE and mesh RMSE, and ConceptFusion reports 3D mIoU 2109.07982 tab. II 2101.06894 tab. 3 2302.07241 tab. 1. These measure different things and should not be compared directly.

Papers

• M2DGR: A Multi-sensor and Multi-scenario SLAM Dataset for Ground Robots: a 36-sequence dataset with a rich sensor suite, and an evaluation of visual and LiDAR SLAM on seven sequences.
• LVI-SAM: Tightly-coupled Lidar-Visual-Inertial Odometry via Smoothing and Mapping: a factor-graph system that fuses a lidar, a camera and an IMU with loop closure; author-reported RMSE against GPS on three self-gathered datasets.
• SuMa++: Efficient LiDAR-based Semantic SLAM: surfel lidar SLAM with semantic labels that remove moving objects from the map and weight the scan alignment; author-reported KITTI results.
• Replica: A Digital Replica of Indoor Spaces: 18 indoor scene reconstructions with HDR textures and semantic labels, built with an in-house SLAM system; no SLAM benchmark is reported.
• GS3LAM: Gaussian Semantic Splatting SLAM: RGB-D SLAM that jointly estimates camera poses and a semantic Gaussian field, with author-reported results on Replica and ScanNet.
• SplaTAM: Splat, Track & Map 3D Gaussians for Dense RGB-D SLAM: dense RGB-D SLAM on 3D Gaussians, with author-reported ATE on Replica, TUM-RGBD, ScanNet and ScanNet++.
• NeuSE: Neural SE(3)-Equivariant Embedding for Long-Term Object-Based Simultaneous Localization and Mapping: latent object codes that add pose constraints and change detection to an existing SLAM system, with author-reported gains on simulated and real sequences.
• Autonomous Driving Small-Scale Cars: A Survey of Recent Development: a survey of small robot-car platforms, with compiled hardware tables and qualitative localisation and mapping comparisons; no experiments.
• A Survey on 3D Gaussian Splatting: a survey of 3DGS that compiles author-reported SLAM, dynamic-scene and avatar results from other papers; it proposes no method of its own.
• Kimera: from SLAM to Spatial Perception with 3D Dynamic Scene Graphs: a visual-inertial mesh SLAM system with object, human, place, room and building layers; author-reported EuRoC accuracy, mesh error and planning times.
• R3LIVE: a Robust, Real-time, RGB-colored, LiDAR-Inertial-Visual state estimation and mapping package: lidar-inertial odometry with a visual step that colours the map by photometric error; author-reported drift on four campus trajectories and relative errors on a seaport sequence.
• NeRF-SLAM: Real-Time Dense Monocular SLAM with Neural Radiance Fields: dense monocular SLAM whose depth covariances weight a radiance-field mapper; author-reported PSNR and depth error on Replica, at about 10 fps.
• ConceptFusion: Open-set Multimodal 3D Mapping: per-point CLIP-aligned features in a 3D map, queried by text, image, click or audio; author-reported UnCoCo results, with no training.
• iSDF: Real-Time Neural Signed Distance Fields for Robot Perception: online neural SDF mapping from posed depth images, with the poses supplied by an external tracker.
• Autonomous Vehicles on the Edge: A Survey on Autonomous Vehicle Racing: a survey of racing software and hardware, with the localisation errors it reports from other work.
• On Evaluation of Embodied Navigation Agents: recommendations for navigation benchmarks, with no experiments of its own. Cited here for its exploration-regime and odometry-statement requests, not for SLAM results.