This paper presents a new implementation of the NeoSLAM algorithm. The proposed version is a complete rewrite of NeoSLAM into a modular architecture using modern frameworks and introduces additional parameters that increase robustness. This work also provides a comparative evaluation between NeoSLAM and RatSLAM using four datasets under different environmental conditions. Experimental results highlight differences in mapping consistency and trajectory reconstruction, demonstrating the effectiveness and practical applicability of the proposed ROS 2–based version. The results indicate that NeoSLAM achieves performance comparable to RatSLAM, with a slight advantage in the evaluated datasets.

Key Contributions

Architectural Refactoring: The original codebase was fully refactored into a modular and extensible architecture, with a redesigned execution pipeline that improves computational efficiency and prevents data loss during real-time operation.
New Tuning Parameters: Additional parameters were introduced to enhance robustness and flexibility. These include image cropping mechanisms to remove irrelevant or harmful regions from the input (as occurs in RatSLAM) and a control parameter to avoid processing temporally redundant frames.
Integration of the modern HTM framework: The legacy NuPIC framework in Python 2.7 was replaced by the modern htm.core \url{https://github.com/htm-community/htm.core} library, implemented in C++17 with Python 3 bindings, ensuring improved performance and maintainability.
Migration to ROS 2: The adoption of htm.core enabled the transition to a ROS 2 Rolling-based architecture, leveraging modern middleware capabilities, improved communication mechanisms, and enhanced scalability.
Experimental Evaluation: NeoSLAM is systematically evaluated against RatSLAM across four distinct datasets, providing a comprehensive comparative analysis.

Temporal Memory state example.

Experimental Results and Parameters

Robotarium Dataset

iRat Australia Dataset

FIU MMC Lake Dataset

FIU BBC Pier Dataset

Authors

Anonymous Authors

Citation


                    @article{anonymous2026BioinspiredSLAM,

                      author = {Anonymous},

                      title = {Bioinspired SLAM: A New Implementation of NeoSLAM and a Comparative Evaluation with RatSLAM.},

                      journal = {International Conference on Automation Science and Engineering},

                      year = {2026},

                }