Open research questions in Robotics and Sensor-Based Localization
68 unresolved questions extracted from the limitations and future-work sections of 408 Robotics and Sensor-Based Localization papers in our library. Each links back to the study that raised it.
What the literature leaves open
Existing methods struggle to generalise to natural settings such as forests due to self-similarity, clutter, and seasonal variability. The lack of distinctive landmarks and strong self-similarity in forests undermine keypoint-based matching and geometric verification.
HOTFLoc++: End-to-End Hierarchical LiDAR Place Recognition, Re-Ranking, and 6-DoF Metric Localisation in Forests · 2026 · DOIRecovering CPP intrinsics from only two views without any known objects remains a challenging problem. The camera and projector parameters are not always known and fixed across reconstruction. Textureless environments often limit the applicability of passive methods.
Navigating dynamic environments with unpredictable obstacles. Reliance on environmental priors, grid maps, or complex cross-modal alignment. Limited detection range of visual cameras and accuracy of LiDAR relying heavily on the localization algorithm.
Self-attention SAC with vision-augmented LiDAR fusion for mapless robot navigation in dynamic environments · 2026 · DOI3D reconstruction of unknown environments is a key application in robotics but is severely limited by the computational and energy capabilities of current aerial platforms.
Quality-Adaptive Multi-UAV 3D Reconstruction with Sparse Workload Redistribution · 2026Yet, the accuracy potential of visual localization has not been systematically investigated against survey-grade demands.
Accuracy potential of visual localization exploiting high-end street-level imagery · 2026Work targeting learning-based representation, lightweight fusion, and realistic deployment benchmarks are warranted.
Further research can be conducted to improve the robustness of the algorithm in different scenarios. The algorithm can be applied to other fields, such as robotics and autonomous vehicles.
Bio-inspired Dyna-LIO: A Frog Vision-based LiDAR SLAM for Engineering Vehicles in Dynamic Environments · 2026 · DOIExisting LiDAR SLAM algorithms are designed for static environments and face challenges in dynamic construction scenarios. There is a need for a novel dynamic LiDAR SLAM algorithm that can achieve significant improvements in localization accuracy in dynamic environments.
Bio-inspired Dyna-LIO: A Frog Vision-based LiDAR SLAM for Engineering Vehicles in Dynamic Environments · 2026 · DOIThe incorrect estimation of SDF values can have a significant impact on the localization and mapping of implicit neural dense SLAM. Implicit neural dense SLAM systems often suffer from lower localization accuracy compared to traditional LiDAR-inertial odometry.
The model still faces challenges with small, long-range objects. Higher-resolution BEV or temporal cues may be needed for future work.
Exploring higher-resolution BEV or temporal cues to improve performance on small, long-range objects. Applying the proposed framework to other fields that require reliable 3-D perception.
The paper suggests that future research should focus on the enhancement of robustness, scalability, and adaptability of SLAM systems. The use of multi-sensor fusion methods and learning-based SLAM approaches is suggested. The paper also suggests that further research should be done on uncertainty estimation and optimizing for resource-constrained platforms.
The paper identifies the gap in SLAM research, including the challenges of dynamic environments, sensor uncertainty, and real-time computation. The gap in the development of robust and scalable SLAM systems is also mentioned.
The model's rotation estimates decrease in accuracy under aggressive motion. Current tests are limited to indoor data. Further outdoor and broader baseline comparisons are needed for comprehensive validation.
Deep Learning for Autonomous UAV Navigation: Multi-Modal Visual-Inertial Pose Estimation in GPS-Denied Environments · 2026 · DOIImproving IMU integration or advanced fusion techniques to enhance rotation estimates under aggressive motion. Conducting further outdoor and broader baseline comparisons for comprehensive validation.
Deep Learning for Autonomous UAV Navigation: Multi-Modal Visual-Inertial Pose Estimation in GPS-Denied Environments · 2026 · DOIThe SEALOC dataset is limited to five benthic reference sites. The geometric registration between visits may have local errors up to about 0.16 m. The method assumes a local tangent plane approximation for coordinate conversion.
Long-term visual localization in dynamic benthic environments: the SEALOC dataset, footprint-based ground truth, and visual place recognition benchmark · 2026 · DOIThe lack of curated datasets for benchmarking long-term visual localization in benthic environments. The need for accurate and efficient methods for ground-truthing visual localization results. The limited understanding of visual place recognition in dynamic benthic environments.
Long-term visual localization in dynamic benthic environments: the SEALOC dataset, footprint-based ground truth, and visual place recognition benchmark · 2026 · DOITo improve the accuracy and efficiency of the framework in more complex and dynamic environments. To explore the use of other semantic segmentation models and their integration with the online A* planner.
Vision-aided online A$$^*$$ path planning for efficient and safe navigation of service robots · 2026 · DOIThe lack of a framework that integrates advanced semantic segmentation with real-time path planning. The need for a cost-effective solution for service robot navigation that can be used on low-cost platforms.
Vision-aided online A$$^*$$ path planning for efficient and safe navigation of service robots · 2026 · DOIAddressing variable lighting and view-dependent effects - Improving pose correction or incremental buildup - Exploring applications in various domains
Existing approaches typically rely on pre-calibrated intrinsics or multi-view self-calibration with known reference objects, which limits their applicability in practical indoor scenarios. Recovering CPP intrinsics from only two views without any known objects remains a challenging problem.
Future research can focus on improving the feed-forward model to handle unseen, unusual scenes. The integration of other sensor modalities, such as lidar or radar, can be explored to further improve the robustness of the system.
MASt3R-Fusion: Integrating feed-forward visual model with IMU, GNSS for high-functionality SLAM · 2026 · DOIThe integration of multi-sensor information in visual SLAM systems has demonstrated remarkable effectiveness, but its application to emerging vision-based methods remains to be explored. The gap lies in the lack of a multi-sensor-assisted visual SLAM framework that can handle low-texture environments and scale ambiguity.
MASt3R-Fusion: Integrating feed-forward visual model with IMU, GNSS for high-functionality SLAM · 2026 · DOIThe method has a relatively low accuracy compared to other approaches, such as Structure-from-Motion. The error may arise from the accuracy of the drone's onboard sensors or the sensitivity of the approach to drone parameters. Additional sensors, such as laser-based altimeters, may be needed to reduce positional errors.
Improving the accuracy of the method by using additional sensors, such as laser-based altimeters. Applying the method to different ecological settings, such as tracking animals at sea or on snow.
Most-cited papers in Robotics and Sensor-Based Localization
- Gaussian Splatting SLAM · 2024 · 427 citations
- SplaTAM: Splat, Track & Map 3D Gaussians for Dense RGB-D SLAM · 2024 · 390 citations
- GS-SLAM: Dense Visual SLAM with 3D Gaussian Splatting · 2024 · 342 citations
- FoundationPose: Unified 6D Pose Estimation and Tracking of Novel Objects · 2024 · 266 citations
- BEVFormer: Learning Bird’s-Eye-View Representation From LiDAR-Camera via Spatiotemporal Transformers · IEEE Transactions on Pattern Analysis and Machine Intelligence · 2024 · 205 citations
- ConceptGraphs: Open-Vocabulary 3D Scene Graphs for Perception and Planning · 2024 · 204 citations
- Photo-SLAM: Real-Time Simultaneous Localization and Photorealistic Mapping for Monocular, Stereo, and RGB-D Cameras · 2024 · 181 citations
- FAST-LIVO2: Fast, Direct LiDAR–Inertial–Visual Odometry · IEEE Transactions on Robotics · 2024 · 140 citations
- Mobile robot localization: Current challenges and future prospective · Computer Science Review · 2024 · 123 citations
- A Survey on Global LiDAR Localization: Challenges, Advances and Open Problems · International Journal of Computer Vision · 2024 · 122 citations
Most recent work
- ZA-SLAM: Leveraging Vision-Language Model for Zero-Shot Acoustic SLAM · 2026
- Joint localization method via affine perception and multimodal fusion-based bird’s-eye-view map registration · Journal of Electronic Imaging · 2026
- Robust Visual-Aided UAV Localization via Sequential Image Correlation and Motion Estimation · Problems of Mechatronics Armament Aviation Safety Engineering · 2026
- Quadruplet-attention transformer for scale-invariant robot place recognition · Expert Systems with Applications · 2026
- Deep learning-based robust optical localization for hypersonic platforms · CEAS Space Journal · 2026
- AI-Driven Autonomous Mobile Robot with Vision-Based SLAM for Intelligent Warehouse Navigation · International Journal of Creative and Open Research in Engineering and Management · 2026
- 2D Mapping System using the Google Cartographer SLAM Algorithm and RPLIDAR · International Journal of Science, Strategic Management and Technology · 2026
- Bio-inspired Dyna-LIO: A Frog Vision-based LiDAR SLAM for Engineering Vehicles in Dynamic Environments · Journal of Bionic Engineering · 2026
- Structure-Aware Indoor RGB-D SLAM via Manhattan-Constrained 2D Gaussian Splatting · CSIAM Transactions on Applied Mathematics · 2026
- Adaptive Terrain-Cognizant Multispectral Vision and Distributed Intelligence for Low-Altitude Unmanned Systems in Yunnan Plateau and Canyon Ecosystems · Advances in Engineering Research Possibilities and Challenges · 2026
Find a gap in your own Robotics and Sensor-Based Localization sub-topic
This page shows what the Robotics and Sensor-Based Localization literature already flags as unresolved. To narrow it to your specific question, run the guided finder — it searches the gap library on demand and checks candidates against 250M+ OpenAlex works.
Open the Research Gap Finder →