Tushar Sandhan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Tushar Sandhan has a total of 52 co-authors. Here are the top co-authors by number of shared publications:
- Fei Cheng
Kyoto University
3 shared publications - Jivnesh Sandhan
Kyoto University Japan
3 shared publications - Laxmidhar Behera
Indian Knowledge System and Mental Health Applications (IKSMHA) Center, IIT Mandi, Mandi, Himachal Pradesh, India
3 shared publications - Om Jee Pandey
Department of Electronics Engineering, IIT (BHU), Varanasi, Uttar Pradesh, India
3 shared publications - Yugo Murawaki
Graduate School of Informatics Kyoto University
3 shared publications - Debojyoti Misra
Indian Institute of Technology Kanpur
2 shared publications
Latest Publications
- 6 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Artificial Intelligence
- Vol. PP  (9) , 1-12
Complex AFNet: A Hybrid Complex-Valued Deep Network for Atrial Fibrillation Detection
Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia, affecting nearly 59 million individuals worldwide by 2019 and significantly increasing stroke risk. This work presents a novel hybrid Complex Convolutional Network for automated AF detection using Gramian Angular Field (GAF) transformations of electrocardiogram signals. The proposed method converts one-dimensional ECG segments into two-dimensional image representations through Gramian Angular Summation Field (GASF) and Gramian Angular Difference Field (GADF) transformations, preserving temporal dependencies as geometric relationships in polar coordinates. Our proposed model treats GASF and GADF as real and imaginary components of complex-valued inputs, enabling phase-aware feature extraction through the proposed hybrid architecture combining realvalued and complex-valued convolutional layers. The network incorporates complex convolutions, modified ReLU activations, and residual connections to capture subtle magnitude-phase interactions critical for detecting AF episodes. Evaluated on the PhysioNet AF dataset containing 1,436 ECG recordings, this model achieves superior performance with 95.63% test accuracy, 96.43% precision, 92.05% recall, and 94.15% F1-score, outperforming state-of-the-art models by +1.52% accuracy and +2.1% F1-score. Interpretability analysis using Grad-CAM and Integrated Gradients reveals class-specific attention patterns, with the model focusing on GADF for normal rhythm detection and GASF for AF classification, demonstrating clinically relevant pattern recognition capabilities of our model. To facilitate reproducibility and further research, the implementation of the proposed C-AFNet and the associated preprocessing and evaluation scripts are publicly available at this link.
- 6 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Artificial Intelligence
- Vol. PP  (9) , 1-12
Complex AFNet: A Hybrid Complex-Valued Deep Network for Atrial Fibrillation Detection
Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia, affecting nearly 59 million individuals worldwide by 2019 and significantly increasing stroke risk. This work presents a novel hybrid Complex Convolutional Network for automated AF detection using Gramian Angular Field (GAF) transformations of electrocardiogram signals. The proposed method converts one-dimensional ECG segments into two-dimensional image representations through Gramian Angular Summation Field (GASF) and Gramian Angular Difference Field (GADF) transformations, preserving temporal dependencies as geometric relationships in polar coordinates. Our proposed model treats GASF and GADF as real and imaginary components of complex-valued inputs, enabling phase-aware feature extraction through the proposed hybrid architecture combining realvalued and complex-valued convolutional layers. The network incorporates complex convolutions, modified ReLU activations, and residual connections to capture subtle magnitude-phase interactions critical for detecting AF episodes. Evaluated on the PhysioNet AF dataset containing 1,436 ECG recordings, this model achieves superior performance with 95.63% test accuracy, 96.43% precision, 92.05% recall, and 94.15% F1-score, outperforming state-of-the-art models by +1.52% accuracy and +2.1% F1-score. Interpretability analysis using Grad-CAM and Integrated Gradients reveals class-specific attention patterns, with the model focusing on GADF for normal rhythm detection and GASF for AF classification, demonstrating clinically relevant pattern recognition capabilities of our model. To facilitate reproducibility and further research, the implementation of the proposed C-AFNet and the associated preprocessing and evaluation scripts are publicly available at this link.
- Open Access
Persona Jailbreaking in Large Language Models
Large Language Models (LLMs) are increasingly deployed in domains such as education, mental health and customer support, where stable and consistent personas are critical for reliability. Yet, existing studies focus on narrative or role-playing tasks and overlook how adversarial conversational history alone can reshape induced personas. Black-box persona manipulation remains unexplored, raising concerns for robustness in realistic interactions. In response, we introduce the task of persona editing, which adversarially steers LLM traits through user-side inputs under a black-box, inference-only setting. To this end, we propose PHISH (Persona Hijacking via Implicit Steering in History), the first framework to expose a new vulnerability in LLM safety that embeds semantically loaded cues into user queries to gradually induce reverse personas. We also define a metric to quantify attack success. Across 3 benchmarks and 8 LLMs, PHISH predictably shifts personas, triggers collateral changes in correlated traits, and exhibits stronger effects in multi-turn settings. In high-risk domains mental health, tutoring, and customer support, PHISH reliably manipulates personas, validated by both human and LLM-as-Judge evaluations. Importantly, PHISH causes only a small reduction in reasoning benchmark performance, leaving overall utility largely intact while still enabling significant persona manipulation. While current guardrails offer partial protection, they remain brittle under sustained attack. Our findings expose new vulnerabilities in personas and highlight the need for context-resilient persona in LLMs. Our codebase and dataset is available at: https://github.com/Jivnesh/PHISH
Open AccessPersona Jailbreaking in Large Language Models
Large Language Models (LLMs) are increasingly deployed in domains such as education, mental health and customer support, where stable and consistent personas are critical for reliability. Yet, existing studies focus on narrative or role-playing tasks and overlook how adversarial conversational history alone can reshape induced personas. Black-box persona manipulation remains unexplored, raising concerns for robustness in realistic interactions. In response, we introduce the task of persona editing, which adversarially steers LLM traits through user-side inputs under a black-box, inference-only setting. To this end, we propose PHISH (Persona Hijacking via Implicit Steering in History), the first framework to expose a new vulnerability in LLM safety that embeds semantically loaded cues into user queries to gradually induce reverse personas. We also define a metric to quantify attack success. Across 3 benchmarks and 8 LLMs, PHISH predictably shifts personas, triggers collateral changes in correlated traits, and exhibits stronger effects in multi-turn settings. In high-risk domains mental health, tutoring, and customer support, PHISH reliably manipulates personas, validated by both human and LLM-as-Judge evaluations. Importantly, PHISH causes only a small reduction in reasoning benchmark performance, leaving overall utility largely intact while still enabling significant persona manipulation. While current guardrails offer partial protection, they remain brittle under sustained attack. Our findings expose new vulnerabilities in personas and highlight the need for context-resilient persona in LLMs. Our codebase and dataset is available at: https://github.com/Jivnesh/PHISH
- 12 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Practice
- Vol. 1, 23-28
Bin-Picking With Category-Agnostic Segmentation for Unreliable Depth Scenarios
Robotic table-top grasping, a.k.a. bin-picking of unknown heterogeneous objects that are put together is a challenging problem due to clutter and uncertainty in depth-data. Out of many types of solutions to this problem, one uses a category-agnostic instance segmentation step before a grasp-planning step to segment out the object boundaries. Recently, various such methods have been proposed that use depth data produced from industrial-grade sensors, as input to their trained deep-learning models to reliably segment out the object instances in the cluttered bin. However, generally, with the commonly used commodity-grade depth sensors, the obtained depth data is noisy, particularly unreliable for non-opaque and thin objects. One other challenge could be time efficiency if the grasp-planning step needs to process each of the segmented objects in the grasp-planning step. To address these challenges, we propose a unified depth-independent CNN design that aims to co-learn category-agnostic instance segmentation, instance- wise grasp-confidence-scores (GCS), and monocular depth estimation, given the RGB image of the scene as the input. The estimated depth is used to detect collisions in grasp pose prediction and to transform the planner grasp pose to the 3D world coordinates. A novel GCS branch is developed into the instance-segmentation model to help filter the detected object instances based on their graspability, avoiding the need of processing each object for grasp planning. A custom-generated synthetic dataset is leveraged to train the proposed CNN architecture. We show through experimentation that our proposed bin-picking system can reliably pick various kinds of unknown objects (opaque, non-opaque, and thin objects) from a clutter of around 20-40 objects.
- 12 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Practice
- Vol. 1, 23-28
Bin-Picking With Category-Agnostic Segmentation for Unreliable Depth Scenarios
Robotic table-top grasping, a.k.a. bin-picking of unknown heterogeneous objects that are put together is a challenging problem due to clutter and uncertainty in depth-data. Out of many types of solutions to this problem, one uses a category-agnostic instance segmentation step before a grasp-planning step to segment out the object boundaries. Recently, various such methods have been proposed that use depth data produced from industrial-grade sensors, as input to their trained deep-learning models to reliably segment out the object instances in the cluttered bin. However, generally, with the commonly used commodity-grade depth sensors, the obtained depth data is noisy, particularly unreliable for non-opaque and thin objects. One other challenge could be time efficiency if the grasp-planning step needs to process each of the segmented objects in the grasp-planning step. To address these challenges, we propose a unified depth-independent CNN design that aims to co-learn category-agnostic instance segmentation, instance- wise grasp-confidence-scores (GCS), and monocular depth estimation, given the RGB image of the scene as the input. The estimated depth is used to detect collisions in grasp pose prediction and to transform the planner grasp pose to the 3D world coordinates. A novel GCS branch is developed into the instance-segmentation model to help filter the detected object instances based on their graspability, avoiding the need of processing each object for grasp planning. A custom-generated synthetic dataset is leveraged to train the proposed CNN architecture. We show through experimentation that our proposed bin-picking system can reliably pick various kinds of unknown objects (opaque, non-opaque, and thin objects) from a clutter of around 20-40 objects.
- Open Access
Vision-Aided Online A* Path Planning for Efficient and Safe Navigation of Service Robots
The deployment of autonomous service robots in human-centric environments is hindered by a critical gap in perception and planning. Traditional navigation systems rely on expensive LiDARs that, while geometrically precise, are semantically unaware, they cannot distinguish a important document on an office floor from a harmless piece of litter, treating both as physically traversable. While advanced semantic segmentation exists, no prior work has successfully integrated this visual intelligence into a real-time path planner that is efficient enough for low-cost, embedded hardware. This paper presents a framework to bridge this gap, delivering context-aware navigation on an affordable robotic platform. Our approach centers on a novel, tight integration of a lightweight perception module with an online A* planner. The perception system employs a semantic segmentation model to identify user-defined visual constraints, enabling the robot to navigate based on contextual importance rather than physical size alone. This adaptability allows an operator to define what is critical for a given task, be it sensitive papers in an office or safety lines in a factory, thus resolving the ambiguity of what to avoid. This semantic perception is seamlessly fused with geometric data. The identified visual constraints are projected as non-geometric obstacles onto a global map that is continuously updated from sensor data, enabling robust navigation through both partially known and unknown environments. We validate our framework through extensive experiments in high-fidelity simulations and on a real-world robotic platform. The results demonstrate robust, real-time performance, proving that a cost-effective robot can safely navigate complex environments while respecting critical visual cues invisible to traditional planners.
Open AccessVision-Aided Online A* Path Planning for Efficient and Safe Navigation of Service Robots
The deployment of autonomous service robots in human-centric environments is hindered by a critical gap in perception and planning. Traditional navigation systems rely on expensive LiDARs that, while geometrically precise, are semantically unaware, they cannot distinguish a important document on an office floor from a harmless piece of litter, treating both as physically traversable. While advanced semantic segmentation exists, no prior work has successfully integrated this visual intelligence into a real-time path planner that is efficient enough for low-cost, embedded hardware. This paper presents a framework to bridge this gap, delivering context-aware navigation on an affordable robotic platform. Our approach centers on a novel, tight integration of a lightweight perception module with an online A* planner. The perception system employs a semantic segmentation model to identify user-defined visual constraints, enabling the robot to navigate based on contextual importance rather than physical size alone. This adaptability allows an operator to define what is critical for a given task, be it sensitive papers in an office or safety lines in a factory, thus resolving the ambiguity of what to avoid. This semantic perception is seamlessly fused with geometric data. The identified visual constraints are projected as non-geometric obstacles onto a global map that is continuously updated from sensor data, enabling robust navigation through both partially known and unknown environments. We validate our framework through extensive experiments in high-fidelity simulations and on a real-world robotic platform. The results demonstrate robust, real-time performance, proving that a cost-effective robot can safely navigate complex environments while respecting critical visual cues invisible to traditional planners.
- Open Access
CAPE: Context-Aware Personality Evaluation Framework for Large Language Models
Psychometric tests, traditionally used to assess humans, are now being applied to Large Language Models (LLMs) to evaluate their behavioral traits. However, existing studies follow a context-free approach, answering each question in isolation to avoid contextual influence. We term this the Disney World test, an artificial setting that ignores real-world applications, where conversational history shapes responses. To bridge this gap, we propose the first Context-Aware Personality Evaluation (CAPE) framework for LLMs, incorporating prior conversational interactions. To thoroughly analyze the influence of context, we introduce novel metrics to quantify the consistency of LLM responses, a fundamental trait in human behavior. Our exhaustive experiments on 7 LLMs reveal that conversational history enhances response consistency via in-context learning but also induces personality shifts, with GPT-3.5-Turbo and GPT-4-Turbo exhibiting extreme deviations. While GPT models are robust to question ordering, Gemini-1.5-Flash and Llama-8B display significant sensitivity. Moreover, GPT models response stem from their intrinsic personality traits as well as prior interactions, whereas Gemini-1.5-Flash and Llama--8B heavily depend on prior interactions. Finally, applying our framework to Role Playing Agents (RPAs) shows context-dependent personality shifts improve response consistency and better align with human judgments. Our code and datasets are publicly available at: https://github.com/jivnesh/CAPE
Open AccessCAPE: Context-Aware Personality Evaluation Framework for Large Language Models
Psychometric tests, traditionally used to assess humans, are now being applied to Large Language Models (LLMs) to evaluate their behavioral traits. However, existing studies follow a context-free approach, answering each question in isolation to avoid contextual influence. We term this the Disney World test, an artificial setting that ignores real-world applications, where conversational history shapes responses. To bridge this gap, we propose the first Context-Aware Personality Evaluation (CAPE) framework for LLMs, incorporating prior conversational interactions. To thoroughly analyze the influence of context, we introduce novel metrics to quantify the consistency of LLM responses, a fundamental trait in human behavior. Our exhaustive experiments on 7 LLMs reveal that conversational history enhances response consistency via in-context learning but also induces personality shifts, with GPT-3.5-Turbo and GPT-4-Turbo exhibiting extreme deviations. While GPT models are robust to question ordering, Gemini-1.5-Flash and Llama-8B display significant sensitivity. Moreover, GPT models response stem from their intrinsic personality traits as well as prior interactions, whereas Gemini-1.5-Flash and Llama--8B heavily depend on prior interactions. Finally, applying our framework to Role Playing Agents (RPAs) shows context-dependent personality shifts improve response consistency and better align with human judgments. Our code and datasets are publicly available at: https://github.com/jivnesh/CAPE
- 29 July 2025
- Published by Springer Nature
- p. 32-46
Boundary Refinement in Abdomen Segmentation Using 3D Inverseform Loss
- 29 July 2025
- Published by Springer Nature
- p. 32-46
Boundary Refinement in Abdomen Segmentation Using 3D Inverseform Loss
- 13 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Reliability
- Vol. 74  (3) , 3514-3528
ENADL: Towards Performance Improvement of IoT Networks Using Deep Learning-Based Node Fault Prediction
The Internet of Things (IoT) has grown explosively with wireless technology integration. Several IoT applications require high data throughput, low data transmission latency, and high data gathering reliability. Since, the IoT network (IoTN) is generally dynamic and utilizes a multi-hop data transmission scheme for such applications, the throughput, latency, and network lifetime tend to degrade as the hops increase. Moreover, IoT devices (IoD) are low-cost, less computationally capable, and battery-limited, further impacting performance. A faulty IoD worsens network lifetime and throughput. Predicting faulty nodes and re-routing data can significantly enhance performance. This work proposes a node fault prediction framework to enhance data routing in dynamic IoTN, maximizing throughput and lifetime. The network is represented as a graph in which the IoD are the nodes. Then a novel deep learning model is proposed utilizing various node and edge features to predict the faulty IoDs. Particularly, the proposed edge and node features-accumulation deep learning (ENADL) method exploits features, such as Euclidean distance between nodes, residual energy level of nodes, and type and number of messages passed between edges to predict the forthcoming faulty IoD. Thereafter, data routing is performed over the updated network topology. Furthermore, to improve the network lifetime, the node's degree and betweenness centrality measures-based energy allocation method is also proposed. Finally, numerical results on simulated and real-field testbeds demonstrate the ENADL method.s effectiveness in predicting faulty nodes and re-routing data packets. This results in maximized network throughput and lifetime as compared to several existing methods.
- 13 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Reliability
- Vol. 74  (3) , 3514-3528
ENADL: Towards Performance Improvement of IoT Networks Using Deep Learning-Based Node Fault Prediction
The Internet of Things (IoT) has grown explosively with wireless technology integration. Several IoT applications require high data throughput, low data transmission latency, and high data gathering reliability. Since, the IoT network (IoTN) is generally dynamic and utilizes a multi-hop data transmission scheme for such applications, the throughput, latency, and network lifetime tend to degrade as the hops increase. Moreover, IoT devices (IoD) are low-cost, less computationally capable, and battery-limited, further impacting performance. A faulty IoD worsens network lifetime and throughput. Predicting faulty nodes and re-routing data can significantly enhance performance. This work proposes a node fault prediction framework to enhance data routing in dynamic IoTN, maximizing throughput and lifetime. The network is represented as a graph in which the IoD are the nodes. Then a novel deep learning model is proposed utilizing various node and edge features to predict the faulty IoDs. Particularly, the proposed edge and node features-accumulation deep learning (ENADL) method exploits features, such as Euclidean distance between nodes, residual energy level of nodes, and type and number of messages passed between edges to predict the forthcoming faulty IoD. Thereafter, data routing is performed over the updated network topology. Furthermore, to improve the network lifetime, the node's degree and betweenness centrality measures-based energy allocation method is also proposed. Finally, numerical results on simulated and real-field testbeds demonstrate the ENADL method.s effectiveness in predicting faulty nodes and re-routing data packets. This results in maximized network throughput and lifetime as compared to several existing methods.
- 7 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Artificial Intelligence
- Vol. 6  (9) , 2486-2497
MalaNet: A Small World Inspired Neural Network for Automated Malaria Diagnosis
In this work, a novel neural network architecture called MalaNet is proposed for the detection and diagnosis of malaria, an infectious disease that poses a major global health challenge. The proposed neural network architecture is inspired by small-world network principles, which generally involve the introduction of new links. A small-world neural network is realized by establishing new connections, thereby reducing the average path length and increasing clustering coefficient. These characteristics are known to enhance interconnectivity and improve feature propagation within the network. In the context of malaria diagnosis, these characteristics of MalaNet can enhance detection accuracy and enable better generalization in scenarios with limited data availability. Broadly, two variants of MalaNet are proposed in this work. First, a small-world-inspired Feed-Forward Neural Network (FNN) is developed for symptom and categorical feature-based diagnosis, providing an accessible solution when blood smear images are unavailable. Subsequently, a small-world-inspired Convolutional Neural Network (CNN) is developed for precise and automated diagnosis when blood smear images are available. Both variants of MalaNet are rigorously validated using the National Institute of Health Malaria dataset, a clinical dataset from Federal Polytechnic Ilaro Medical Centre, Nigeria, and the APTOS dataset. Comparative results against several state-of-the-art neural network models in the literature demonstrate MalaNet’s superior performance, generalization capability, and computational efficiency. The small-world neural network architecture proposed in this work enhances feature learning, diagnostic accuracy, and adaptability in limited-data and resource-constrained settings, motivating its application in disease diagnosis where timely and accurate results are critical.
- 7 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Artificial Intelligence
- Vol. 6  (9) , 2486-2497
MalaNet: A Small World Inspired Neural Network for Automated Malaria Diagnosis
In this work, a novel neural network architecture called MalaNet is proposed for the detection and diagnosis of malaria, an infectious disease that poses a major global health challenge. The proposed neural network architecture is inspired by small-world network principles, which generally involve the introduction of new links. A small-world neural network is realized by establishing new connections, thereby reducing the average path length and increasing clustering coefficient. These characteristics are known to enhance interconnectivity and improve feature propagation within the network. In the context of malaria diagnosis, these characteristics of MalaNet can enhance detection accuracy and enable better generalization in scenarios with limited data availability. Broadly, two variants of MalaNet are proposed in this work. First, a small-world-inspired Feed-Forward Neural Network (FNN) is developed for symptom and categorical feature-based diagnosis, providing an accessible solution when blood smear images are unavailable. Subsequently, a small-world-inspired Convolutional Neural Network (CNN) is developed for precise and automated diagnosis when blood smear images are available. Both variants of MalaNet are rigorously validated using the National Institute of Health Malaria dataset, a clinical dataset from Federal Polytechnic Ilaro Medical Centre, Nigeria, and the APTOS dataset. Comparative results against several state-of-the-art neural network models in the literature demonstrate MalaNet’s superior performance, generalization capability, and computational efficiency. The small-world neural network architecture proposed in this work enhances feature learning, diagnostic accuracy, and adaptability in limited-data and resource-constrained settings, motivating its application in disease diagnosis where timely and accurate results are critical.
- 10 February 2025
- Published by Emerald Publishing in International Journal of Intelligent Unmanned Systems
- Vol. 13  (1) , 92-128
A cyber-physical system based unmanned ground vehicles for safety inspection and rescue support in an underground mine
Purpose: The unmanned ground vehicle (UGV) described in this manuscript is a robot designed by the authors to map the underground mine environments. The UGV works to develop a computational intelligence-based cyber-physical system (CPS)-based analytical framework for mining operations. The UGV demonstrated excellent semi-autonomous navigation capabilities in the absence of GNSS signals. The UGV has a suite that works in unison to provide relevant information. These sensors are integrated to form a robust sensor fusion-based architecture, creating a CPS with a wide range of capabilities such as data acquisition and navigation in challenging underground environments. UGVs can be used to enhance the efficacy of safety inspections, rescue during underground emergencies and assist miners in hazardous conditions. Design/methodology/approach: In this research, an UGV was constructed whose operations are enabled by sensors including a D415i Red Blue Green (RGB) depth camera, a LiDAR, a FLIR C5 infrared camera and smart air quality sensors. This sensor fusion-based architecture forms a CPS. Data obtained remotely are processed by deep learning algorithms to achieve overall capabilities such as real-time image analysis for miner identification, object detection, posture analysis and identifying threats of roof falls and overhangs. Simultaneous localization and mapping (SLAM) algorithms create a 3D map, facilitate autonomous navigation and build a decision support system for delivering mine rescue support. Findings: The aim of this study is to include this capacity in training situations when it has been validated and authorized by the Directorate General of Mines Safety (DGMS) Indian government regulatory agency for safety in mines and oil fields. The longwall demo mine, at IIT (ISM) is being used as the site of the first operations. Once approved by the respective enforcement agencies, this technology and the accompanying rescue and training process can be used in underground operations. Originality/value: In fact, this paper is the first attempt at remotely operated UGVs based on CPSs, the CPS–UGV in Indian mine conditions, so as to revolutionize Indian mines based on the idea of Industry 4.0.
- 10 February 2025
- Published by Emerald Publishing in International Journal of Intelligent Unmanned Systems
- Vol. 13  (1) , 92-128
A cyber-physical system based unmanned ground vehicles for safety inspection and rescue support in an underground mine
Purpose: The unmanned ground vehicle (UGV) described in this manuscript is a robot designed by the authors to map the underground mine environments. The UGV works to develop a computational intelligence-based cyber-physical system (CPS)-based analytical framework for mining operations. The UGV demonstrated excellent semi-autonomous navigation capabilities in the absence of GNSS signals. The UGV has a suite that works in unison to provide relevant information. These sensors are integrated to form a robust sensor fusion-based architecture, creating a CPS with a wide range of capabilities such as data acquisition and navigation in challenging underground environments. UGVs can be used to enhance the efficacy of safety inspections, rescue during underground emergencies and assist miners in hazardous conditions. Design/methodology/approach: In this research, an UGV was constructed whose operations are enabled by sensors including a D415i Red Blue Green (RGB) depth camera, a LiDAR, a FLIR C5 infrared camera and smart air quality sensors. This sensor fusion-based architecture forms a CPS. Data obtained remotely are processed by deep learning algorithms to achieve overall capabilities such as real-time image analysis for miner identification, object detection, posture analysis and identifying threats of roof falls and overhangs. Simultaneous localization and mapping (SLAM) algorithms create a 3D map, facilitate autonomous navigation and build a decision support system for delivering mine rescue support. Findings: The aim of this study is to include this capacity in training situations when it has been validated and authorized by the Directorate General of Mines Safety (DGMS) Indian government regulatory agency for safety in mines and oil fields. The longwall demo mine, at IIT (ISM) is being used as the site of the first operations. Once approved by the respective enforcement agencies, this technology and the accompanying rescue and training process can be used in underground operations. Originality/value: In fact, this paper is the first attempt at remotely operated UGVs based on CPSs, the CPS–UGV in Indian mine conditions, so as to revolutionize Indian mines based on the idea of Industry 4.0.