Chandan Sheikder
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Chandan Sheikder has a total of 31 co-authors. Here are the top co-authors by number of shared publications:
- Weimin Zhang
School of Mechatronical Engineering, Beijing Institute of Technology, Beijing 100081, China
10 shared publications Xiaopeng Chen
7 shared publications- Xiaopeng Chen
Beijing Institute of Technology, Beijing, China
7 shared publications Fangxing Li
Beijing Institute of Technology
6 shared publications- Xiaohai He
School of Mechatronical Engineering, Beijing Institute of Technology, Beijing 100081, China
6 shared publications Xinyan Tan
Beijing Institute of Technology
6 shared publications
Latest Publications
- Open Access
- 7 October 2026
- Published by Elsevier in SSRN Electronic Journal
Hybrid Jetson Orin Nano and Intel Loihi 2 for Low-Power Robot Navigation
Small autonomous robots require multi-sensor fusion for navigation in GPS-denied environments, but are constrained by power and memory limits. This work introduces a hybrid architecture combining an NVIDIA Jetson Orin Nano with an Intel Loihi 2 neuromorphic processor to offload event-driven sensor fusion algorithms to spiking neural networks. The system reduces total power from 60.9 W to 22.7 W (62.7%) and cuts fusion-core power from 21.3 W to 0.145 W (99.3%), while memory usage drops from 2.2 GB to 0.267 MB. Field tests on a 2 kg quadcopter demonstrate mission duration extending from 2.1 to 4.5 hours. We identify three limitations: memory exhaustion, communication latency, and hardware availability, and implement engineering solutions including a custom PCIe carrier and an incremental map compression algorithm. Physical hardware testing confirms this compression eliminates out-of-memory crashes entirely (0% failure rate) while capping memory at 5.04 GB. Furthermore, standardized benchmarking proves the hybrid system outperforms optimized GPU baselines, achieving 84.3 FPS with an energy efficiency of 0.0126 Joules/frame, enabling highly stable, real-time navigation for resource-constrained platforms.
- 7 October 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessHybrid Jetson Orin Nano and Intel Loihi 2 for Low-Power Robot Navigation
Small autonomous robots require multi-sensor fusion for navigation in GPS-denied environments, but are constrained by power and memory limits. This work introduces a hybrid architecture combining an NVIDIA Jetson Orin Nano with an Intel Loihi 2 neuromorphic processor to offload event-driven sensor fusion algorithms to spiking neural networks. The system reduces total power from 60.9 W to 22.7 W (62.7%) and cuts fusion-core power from 21.3 W to 0.145 W (99.3%), while memory usage drops from 2.2 GB to 0.267 MB. Field tests on a 2 kg quadcopter demonstrate mission duration extending from 2.1 to 4.5 hours. We identify three limitations: memory exhaustion, communication latency, and hardware availability, and implement engineering solutions including a custom PCIe carrier and an incremental map compression algorithm. Physical hardware testing confirms this compression eliminates out-of-memory crashes entirely (0% failure rate) while capping memory at 5.04 GB. Furthermore, standardized benchmarking proves the hybrid system outperforms optimized GPU baselines, achieving 84.3 FPS with an energy efficiency of 0.0126 Joules/frame, enabling highly stable, real-time navigation for resource-constrained platforms.
- Open Access
- 12 August 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (16) , 8028
LV-GS SLAM: A Decoupled LiDAR–Visual 3D Gaussian Splatting SLAM in a Large-Scale Environment
3D Gaussian Splatting (3DGS) has gained prominence in autonomous driving and robotics for its rendering efficiency and high-fidelity reconstruction capabilities. However, incremental 3DGS map construction at large scales remains challenging due to sensor sparsity and computational constraints. In this paper, we propose LV-GS SLAM, a novel system that integrates LiDAR and visual data for incremental, large-scale reconstruction with real-time tracking. This system provides fast and robust pose estimation while also enabling photorealistic rendering. We first employ a LiDAR odometry frontend that processes 30 Hz LiDAR inputs to provide robust initial poses. In our implementation, the complete LiDAR tracking pipeline runs at 15–19 Hz, while the mapping module performs incremental optimization on selected keyframes. To address the sparsity-induced surface discontinuity in conventional LiDAR-based reconstruction, we propose a novel depth propagation approach that initializes 3D Gaussian primitives using dense depth maps, achieving faster PSNR convergence with 9× fewer optimization iterations compared to direct LiDAR initialization. Furthermore, we develop a keyframe-based submap management framework that dynamically adjusts memory allocation based on both primitive density and inter-frame overlap ratio, effectively preventing GPU memory overflow. Our system has been validated on the KITTI dataset, achieving superior rendering quality compared with representative reproducible baselines. We further validate the robustness of the system on a quadruped robot platform, demonstrating satisfactory performance in both pose estimation and high-fidelity reconstruction.
- 12 August 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (16) , 8028
Open AccessLV-GS SLAM: A Decoupled LiDAR–Visual 3D Gaussian Splatting SLAM in a Large-Scale Environment
3D Gaussian Splatting (3DGS) has gained prominence in autonomous driving and robotics for its rendering efficiency and high-fidelity reconstruction capabilities. However, incremental 3DGS map construction at large scales remains challenging due to sensor sparsity and computational constraints. In this paper, we propose LV-GS SLAM, a novel system that integrates LiDAR and visual data for incremental, large-scale reconstruction with real-time tracking. This system provides fast and robust pose estimation while also enabling photorealistic rendering. We first employ a LiDAR odometry frontend that processes 30 Hz LiDAR inputs to provide robust initial poses. In our implementation, the complete LiDAR tracking pipeline runs at 15–19 Hz, while the mapping module performs incremental optimization on selected keyframes. To address the sparsity-induced surface discontinuity in conventional LiDAR-based reconstruction, we propose a novel depth propagation approach that initializes 3D Gaussian primitives using dense depth maps, achieving faster PSNR convergence with 9× fewer optimization iterations compared to direct LiDAR initialization. Furthermore, we develop a keyframe-based submap management framework that dynamically adjusts memory allocation based on both primitive density and inter-frame overlap ratio, effectively preventing GPU memory overflow. Our system has been validated on the KITTI dataset, achieving superior rendering quality compared with representative reproducible baselines. We further validate the robustness of the system on a quadruped robot platform, demonstrating satisfactory performance in both pose estimation and high-fidelity reconstruction.
- Open Access
- 30 June 2026
- Published by MDPI AG in Astronautics
- Vol. 1 (3) , 11
Hybrid Neuromorphic Edge Computing and Quantum Cloud Optimization for Martian Swarm Robot Survival and Map Recovery
Martian dust storms cut off communication and break standard robot navigation. We built a hybrid system that keeps robot swarms alive during these blackouts and recovers their data quickly. Our rovers use Spiking Neural Networks (SNNs) on their own edge processors to navigate without a signal. Once the storm passes, we use the Quantum Approximate Optimization Algorithm (QAOA) on a cloud platform to merge the fragmented maps the rovers collected while they were offline. We tested this system in a Robot Operating System 2 (ROS 2) and Gazebo environment using a simulated 10-rover Martian deployment. During the simulated blackout, our SNN edge navigation achieved a 92.0% survival rate, outperforming traditional planners like Dynamic Window Approach (DWA) (29.0%) and Timed Elastic Band (TEB) (24.3%). The neuromorphic approach also reduced overall system power consumption by 80.0% compared to a traditional unoptimized Graphics Processing Unit (GPU)-based Simultaneous Localization and Mapping (SLAM) baseline. For the map recovery phase, our simulated QAOA proof-of-concept evaluated the map constraints in just 1.2 ms, compared to 50.0 ms for a classical Generalized Iterative Closest Point (G-ICP) and g2o pose-graph approach. Despite the noisy sensor data collected during the blackout, the final quantum-stitched map achieved an 8.54 cm Root Mean Square Error (RMSE). These results show that combining edge-based neuromorphic processing with quantum cloud computing secures swarm survival and accelerates post-disaster data recovery for deep-space missions.
- 30 June 2026
- Published by MDPI AG in Astronautics
- Vol. 1 (3) , 11
Open AccessHybrid Neuromorphic Edge Computing and Quantum Cloud Optimization for Martian Swarm Robot Survival and Map Recovery
Martian dust storms cut off communication and break standard robot navigation. We built a hybrid system that keeps robot swarms alive during these blackouts and recovers their data quickly. Our rovers use Spiking Neural Networks (SNNs) on their own edge processors to navigate without a signal. Once the storm passes, we use the Quantum Approximate Optimization Algorithm (QAOA) on a cloud platform to merge the fragmented maps the rovers collected while they were offline. We tested this system in a Robot Operating System 2 (ROS 2) and Gazebo environment using a simulated 10-rover Martian deployment. During the simulated blackout, our SNN edge navigation achieved a 92.0% survival rate, outperforming traditional planners like Dynamic Window Approach (DWA) (29.0%) and Timed Elastic Band (TEB) (24.3%). The neuromorphic approach also reduced overall system power consumption by 80.0% compared to a traditional unoptimized Graphics Processing Unit (GPU)-based Simultaneous Localization and Mapping (SLAM) baseline. For the map recovery phase, our simulated QAOA proof-of-concept evaluated the map constraints in just 1.2 ms, compared to 50.0 ms for a classical Generalized Iterative Closest Point (G-ICP) and g2o pose-graph approach. Despite the noisy sensor data collected during the blackout, the final quantum-stitched map achieved an 8.54 cm Root Mean Square Error (RMSE). These results show that combining edge-based neuromorphic processing with quantum cloud computing secures swarm survival and accelerates post-disaster data recovery for deep-space missions.
- Open Access
Bio-Inspired Swarm Navigation on Resource-Constrained Robots for GPS-Denied Environments
Experimental validation delivers five quantified outcomes. First, optical pheromone detection achieves 88.7% ± 0.6% accuracy (n = 150, 95% CI), and the dual-modality combined channel achieves 86.1% ± 0.9% (n = 200), with robustness confirmed under 50/60 Hz flicker interference, rapid 200–1200 lux light transitions (485 ms settling), and reflective glare spots. Second, the MQ-135 chemical channel calibration holds R2 ≥ 0.999 across temperatures of 15–35 °C and humidity of 30–90%, with maximum voltage drift of 0.093 V at the highest temperature. Third, 3.2× CNN inference speedup through 8-bit quantisation runs at 15 FPS within 1.8 W. Fourth, peripheral subsystems draw a measured mean of 1.19 W ± 0.02 W (n = 60, 95% CI); the complete per-robot system, including the Jetson Orin Nano compute rail, draws 6.15 W ± 0.09 W, enabling six-hour missions from the 55.08 Wh battery. Fifth, localisation across ten trials yields the mean position error 0.074 m and RMSE 0.081 m with 97.5% map coverage; physical multi-robot tests with 5–8 robots confirm map convergence times of 120–210 steps with collision rates below 0.042 per robot per step. To the best of our knowledge, no prior physical swarm platform has simultaneously demonstrated this combination of capabilities under comparable constraints.
Open AccessBio-Inspired Swarm Navigation on Resource-Constrained Robots for GPS-Denied Environments
Experimental validation delivers five quantified outcomes. First, optical pheromone detection achieves 88.7% ± 0.6% accuracy (n = 150, 95% CI), and the dual-modality combined channel achieves 86.1% ± 0.9% (n = 200), with robustness confirmed under 50/60 Hz flicker interference, rapid 200–1200 lux light transitions (485 ms settling), and reflective glare spots. Second, the MQ-135 chemical channel calibration holds R2 ≥ 0.999 across temperatures of 15–35 °C and humidity of 30–90%, with maximum voltage drift of 0.093 V at the highest temperature. Third, 3.2× CNN inference speedup through 8-bit quantisation runs at 15 FPS within 1.8 W. Fourth, peripheral subsystems draw a measured mean of 1.19 W ± 0.02 W (n = 60, 95% CI); the complete per-robot system, including the Jetson Orin Nano compute rail, draws 6.15 W ± 0.09 W, enabling six-hour missions from the 55.08 Wh battery. Fifth, localisation across ten trials yields the mean position error 0.074 m and RMSE 0.081 m with 97.5% map coverage; physical multi-robot tests with 5–8 robots confirm map convergence times of 120–210 steps with collision rates below 0.042 per robot per step. To the best of our knowledge, no prior physical swarm platform has simultaneously demonstrated this combination of capabilities under comparable constraints.
- 16 April 2026
- Published by Wiley in Journal of Field Robotics
- Vol. 43 (6) , 3752-3782
Soft Computing Techniques Applied to Adaptive Hybrid Navigation Methods for Tethered Robots in Dynamic Environments
Tethered robots face unique constraints: the physical tether can cause mission failure through snagging or excessive tension. Existing methods either require global maps (Tethered A*) or lack tether awareness (Bug algorithms). We present the first hybrid framework combining reactive obstacle avoidance with predictive tether management. Our system integrates (1) a tether‐aware cost function for 3–5 step snag prediction, (2) fuzzy logic for dynamic tension adaptation, and (3) genetic algorithms for tether routing optimization. Validation on a humanoid robot (ATLAS‐T) with instrumented sensors demonstrates 52% fewer snag incidents versus state‐of‐the‐art methods (Tethered A*, TSLAM), 94.7% success in tether‐critical scenarios, and 43% better odometry‐drift resilience. Real‐world experiments confirm practical viability.
- 16 April 2026
- Published by Wiley in Journal of Field Robotics
- Vol. 43 (6) , 3752-3782
Soft Computing Techniques Applied to Adaptive Hybrid Navigation Methods for Tethered Robots in Dynamic Environments
Tethered robots face unique constraints: the physical tether can cause mission failure through snagging or excessive tension. Existing methods either require global maps (Tethered A*) or lack tether awareness (Bug algorithms). We present the first hybrid framework combining reactive obstacle avoidance with predictive tether management. Our system integrates (1) a tether‐aware cost function for 3–5 step snag prediction, (2) fuzzy logic for dynamic tension adaptation, and (3) genetic algorithms for tether routing optimization. Validation on a humanoid robot (ATLAS‐T) with instrumented sensors demonstrates 52% fewer snag incidents versus state‐of‐the‐art methods (Tethered A*, TSLAM), 94.7% success in tether‐critical scenarios, and 43% better odometry‐drift resilience. Real‐world experiments confirm practical viability.
- 18 February 2026
- Published by American Institute of Aeronautics and Astronautics (AIAA) in Journal of Aerospace Information Systems
- p. 1-20
Autonomous Artificial Intelligence for Extreme Radiation Space Missions
Deep-space missions to icy moons like Europa require AI systems capable of surviving extreme radiation (150 krad/year), operating autonomously under hour-long communication delays, and enforcing planetary protection protocols. This study introduces an integrated framework combining 1) radiation-hardened hybrid circuits reducing hardware failures by 61.6% ([ ]) and single-event upsets by 85.5% under simulated Jovian radiation, 2) hierarchical reinforcement learning reducing Earth dependence by 40% and mission planning time by 50%, and 3) context-aware ethical protocols enforcing COSPAR Category IVc contamination limits ([ ]), reducing risks by 78% while maintaining 81% data efficiency. Validated against NASA’s Europa Clipper parameters and Mars rover analogs, the framework demonstrates scalability to Titan and exoplanet missions. Despite 133% higher unit costs, radiation-hardened components require 65% fewer replacements over 5-year missions. This work bridges critical gaps in deep-space AI, enabling sustainable exploration of ocean worlds.
- 18 February 2026
- Published by American Institute of Aeronautics and Astronautics (AIAA) in Journal of Aerospace Information Systems
- p. 1-20
Autonomous Artificial Intelligence for Extreme Radiation Space Missions
Deep-space missions to icy moons like Europa require AI systems capable of surviving extreme radiation (150 krad/year), operating autonomously under hour-long communication delays, and enforcing planetary protection protocols. This study introduces an integrated framework combining 1) radiation-hardened hybrid circuits reducing hardware failures by 61.6% ([ ]) and single-event upsets by 85.5% under simulated Jovian radiation, 2) hierarchical reinforcement learning reducing Earth dependence by 40% and mission planning time by 50%, and 3) context-aware ethical protocols enforcing COSPAR Category IVc contamination limits ([ ]), reducing risks by 78% while maintaining 81% data efficiency. Validated against NASA’s Europa Clipper parameters and Mars rover analogs, the framework demonstrates scalability to Titan and exoplanet missions. Despite 133% higher unit costs, radiation-hardened components require 65% fewer replacements over 5-year missions. This work bridges critical gaps in deep-space AI, enabling sustainable exploration of ocean worlds.
- Open Access
HV-LIOM: Adaptive Hash-Voxel LiDAR–Inertial SLAM with Multi-Resolution Relocalization and Reinforcement Learning for Autonomous Exploration
This paper presents HV-LIOM (Adaptive Hash-Voxel LiDAR–Inertial Odometry and Mapping), a unified LiDAR–inertial SLAM and autonomous exploration framework for real-time 3D mapping in dynamic, GNSS-denied environments. We propose an adaptive hash-voxel mapping scheme that improves memory efficiency and real-time state estimation by subdividing voxels according to local geometric complexity and point density. To enhance robustness to poor initialization, we introduce a multi-resolution relocalization strategy that enables reliable localization against a prior map under large initial pose errors. A learning-based loop-closure module further detects revisited places and injects global constraints, while global pose-graph optimization maintains long-term map consistency. For autonomous exploration, we integrate a Soft Actor–Critic (SAC) policy that selects informative navigation targets online, improving exploration efficiency in unknown scenes. We evaluate HV-LIOM on public datasets (Hilti and NCLT) and a custom mobile robot platform. Results show that HV-LIOM improves absolute pose accuracy by up to 15.2% over FAST-LIO2 in indoor settings and by 7.6% in large-scale outdoor scenarios. The learned exploration policy achieves comparable or superior area coverage with reduced travel distance and exploration time relative to sampling-based and learning-based baselines.
Open AccessHV-LIOM: Adaptive Hash-Voxel LiDAR–Inertial SLAM with Multi-Resolution Relocalization and Reinforcement Learning for Autonomous Exploration
This paper presents HV-LIOM (Adaptive Hash-Voxel LiDAR–Inertial Odometry and Mapping), a unified LiDAR–inertial SLAM and autonomous exploration framework for real-time 3D mapping in dynamic, GNSS-denied environments. We propose an adaptive hash-voxel mapping scheme that improves memory efficiency and real-time state estimation by subdividing voxels according to local geometric complexity and point density. To enhance robustness to poor initialization, we introduce a multi-resolution relocalization strategy that enables reliable localization against a prior map under large initial pose errors. A learning-based loop-closure module further detects revisited places and injects global constraints, while global pose-graph optimization maintains long-term map consistency. For autonomous exploration, we integrate a Soft Actor–Critic (SAC) policy that selects informative navigation targets online, improving exploration efficiency in unknown scenes. We evaluate HV-LIOM on public datasets (Hilti and NCLT) and a custom mobile robot platform. Results show that HV-LIOM improves absolute pose accuracy by up to 15.2% over FAST-LIO2 in indoor settings and by 7.6% in large-scale outdoor scenarios. The learned exploration policy achieves comparable or superior area coverage with reduced travel distance and exploration time relative to sampling-based and learning-based baselines.
- Open Access
- 3 November 2025
- Published by Elsevier in SSRN Electronic Journal
A Neuromorphic Framework for Bio-Inspired Navigation in Autonomous Robots
- 3 November 2025
- Published by Elsevier in SSRN Electronic Journal
Open AccessA Neuromorphic Framework for Bio-Inspired Navigation in Autonomous Robots
- Open Access
Marine-Inspired Multimodal Sensor Fusion and Neuromorphic Processing for Autonomous Navigation in Unstructured Subaquatic Environments
Autonomous navigation in GPS-denied, unstructured environments such as murky waters or complex seabeds remains a formidable challenge for robotic systems, primarily due to sensory degradation and the computational inefficiency of conventional algorithms. Drawing inspiration from the robust navigation strategies of marine species such as the sea turtle’s quantum-assisted magnetoreception, the octopus’s tactile-chemotactic integration, and the jellyfish’s energy-efficient flow sensing this study introduces a novel neuromorphic framework for resilient robotic navigation, fundamentally based on the co-design of marine-inspired sensors and event-based neuromorphic processors. Current systems lack the dynamic, context-aware multisensory fusion observed in these animals, leading to heightened susceptibility to sensor failures and environmental perturbations, as well as high power consumption. This work directly bridges this gap. Our primary contribution is a hybrid sensor fusion model that co-designs advanced sensing replicating the distributed neural processing of cephalopods and the quantum coherence mechanisms of migratory marine fauna with a neuromorphic processing backbone. Enabling real-time, energy-efficient path integration and cognitive mapping without reliance on traditional methods. This proposed framework has the potential to significantly enhance navigational robustness by overcoming the limitations of state-of-the-art solutions. The findings suggest the potential of marine bio-inspired design for advancing autonomous systems in critical applications such as deep-sea exploration, environmental monitoring, and underwater infrastructure inspection.
Open AccessMarine-Inspired Multimodal Sensor Fusion and Neuromorphic Processing for Autonomous Navigation in Unstructured Subaquatic Environments
Autonomous navigation in GPS-denied, unstructured environments such as murky waters or complex seabeds remains a formidable challenge for robotic systems, primarily due to sensory degradation and the computational inefficiency of conventional algorithms. Drawing inspiration from the robust navigation strategies of marine species such as the sea turtle’s quantum-assisted magnetoreception, the octopus’s tactile-chemotactic integration, and the jellyfish’s energy-efficient flow sensing this study introduces a novel neuromorphic framework for resilient robotic navigation, fundamentally based on the co-design of marine-inspired sensors and event-based neuromorphic processors. Current systems lack the dynamic, context-aware multisensory fusion observed in these animals, leading to heightened susceptibility to sensor failures and environmental perturbations, as well as high power consumption. This work directly bridges this gap. Our primary contribution is a hybrid sensor fusion model that co-designs advanced sensing replicating the distributed neural processing of cephalopods and the quantum coherence mechanisms of migratory marine fauna with a neuromorphic processing backbone. Enabling real-time, energy-efficient path integration and cognitive mapping without reliance on traditional methods. This proposed framework has the potential to significantly enhance navigational robustness by overcoming the limitations of state-of-the-art solutions. The findings suggest the potential of marine bio-inspired design for advancing autonomous systems in critical applications such as deep-sea exploration, environmental monitoring, and underwater infrastructure inspection.