Narasimhan Sundararajan

In this paper, a generalizable Decentralized sequential multitask Assignment Learning Spiking neural network (DeALS) approach is presented for solving a Perimeter Defense Problem (PDP). In PDP, a team of defenders is considered to operate on the perimeter of the territory to be protected from incoming intruders. The territory is divided into multiple angular segments. A multitask assignment Spiking Neural Network (SNN) is presented that allocates to defenders, appropriate perimeter segments corresponding to the intruder entry points, in a decentralized manner. The SNN is trained in a supervised manner based on a ground truth developed from a centralized solution for PDP. Due to the decentralized formulation, the SNN trained for one defender can be used to obtain the assignments for other defenders without the need for re-training. A consensus algorithm is also used to resolve any conflicts present in the defenders' assigned trajectories. Detailed performance evaluation studies show that the proposed DeALS approach produces an approximate 11% higher success rate in defending the perimeter when compared with other existing state-of the-art decentralized sequential assignment strategies. Detailed Ablation studies indicate that DeALS is easily generalizable with good performance when there are large variations in the intruder velocities, shape of the territory, size of the defender team, and the area to be protected.

Underwater images suffer severe degradation due to wavelength-dependent attenuation, scattering, and illumination non-uniformity that vary across water types and depths. We propose an unsupervised Domain-Invariant Visual Enhancement and Restoration (DIVER) framework that integrates empirical correction with physics-guided modeling for robust underwater image enhancement. DIVER first applies either IlluminateNet for adaptive luminance enhancement or a Spectral Equalization Filter for spectral normalization. An Adaptive Optical Correction Module then refines hue and contrast using channel-adaptive filtering, while Hydro-OpticNet employs physics-constrained learning to compensate for backscatter and wavelength-dependent attenuation. The parameters of IlluminateNet and Hydro-OpticNet are optimized via unsupervised learning using a composite loss function. DIVER is evaluated on eight diverse datasets covering shallow, deep, and highly turbid environments, including both naturally low-light and artificially illuminated scenes, using reference and non-reference metrics. While state-of-the-art methods such as WaterNet, UDNet, and Phaseformer perform reasonably in shallow water, their performance degrades in deep, unevenly illuminated, or artificially lit conditions. In contrast, DIVER consistently achieves best or near-best performance across all datasets, demonstrating strong domain-invariant capability. DIVER yields at least a 9% improvement over SOTA methods in UCIQE. On the low-light SeaThru dataset, where color-palette references enable direct evaluation of color restoration, DIVER achieves at least a 4.9% reduction in GPMAE compared to existing methods. Beyond visual quality, DIVER also improves robotic perception by enhancing ORB-based keypoint repeatability and matching performance, confirming its robustness across diverse underwater environments.

This paper presents the development of a virtual traffic island-based approach for Urban Airspace Management (UAM) system to resolve simultaneous conflicts. Increasing usage of Unmanned Aerial Vehicle (UAV)-based transportation in urban environments requires a high-density conflict management solution that is both safe and implementable in real time. Existing state-of-the-art methods, including Control Barrier Function (CBF)-based approaches, can guarantee safety, but are not scalable when the number of UAVs is larger. In high-density traffic scenario, the simultaneous existence of multiple conflicts makes real-time execution of conflict resolution a challenging problem. This paper presents a scalable, safe, and implementable dynamic conflict resolution approach called the Virtual Traffic Island (VTI) approach for urban airspace management. The novel VTI approach employs an evolving virtual traffic island with changing radii to handle multiple conflicts simultaneously. In the proposed approach, since each UAV in conflict avoids a single virtual obstacle by solving a CBF-based quadratic programming (QP) problem, the computational requirement for each UAV remains the same, resulting in its easy scalability and safety. The proposed method has been validated for dynamic conflict resolution of a growing number of UAVs in the urban airspace. To validate the implementation aspects of the proposed approach in real time, a software-in-the-loop (SITL) simulation has also been presented. Furthermore, a comparison with other well-known existing methods shows that the novel VTI outperforms these in terms of UAVs reaching their goals and reduces computational complexity, resulting in an easier implementation.

In this paper, a Centralized Conflict-free Assignment Learning Spiking neural network (C2ALS) is formulated for solving a Perimeter Defense Problem (PDP). Here, the region between the perimeter and the sensing range of the defender is divided into two layers. The outermost layer, in which the defender can only sense the intruders, is termed the sensing layer. The layer closest to the perimeter in which the defenders can sense and capture the intruders is termed the capture layer. Both layers are further divided into angular segments with respect to the center of the area to be protected. The spatiotemporal movements of both the defenders and intruders in these segments are converted into spikes and given as input to a Spiking Neural Network (SNN). The SNN is trained in a supervised manner to learn the intruder capture assignments of a defender in terms of these segments. Here, the desired assignments required for training a defender are referred to as greedy assignments because while generating them, priority is given to those segments that are relatively closer to the defender’s actual location. Inhibitory connections are used in the SNN architecture to obtain assignments without conflicts. Detailed performance study results show that the proposed C2ALS approach outperforms the other existing centralized optimization-based solutions for PDP by a 17% increase in the success rates of capturing the intruders. Also, C2ALS is the first centralized spike-based learning solution for PDP, capable of generating conflict-free assignments while ensuring robust intruder assignments with minimal defender movement.

The safety and efficiency of an Autonomous Vehicle (AV) manoeuvre planning heavily depend on the future trajectories of surrounding vehicles. If an AV can predict its surrounding vehicles’ future trajectories, it can make safe and efficient manoeuvre decisions. In this paper, we present a Deep Spatio-Temporal Context-Aware decision Network (DST-CAN) for predictive manoeuvre decisions for AVs on highways. DST-CAN has two main components, namely spatio-temporal context-aware map generator and predictive manoeuvre decisions engine. DST-CAN employ a memory neuron network to predict the future trajectories of its surrounding vehicles. Using look-ahead prediction and past actual trajectories, a spatio-temporal context-aware probability occupancy map is generated. These context-aware maps as input to a decision engine generate a safe and efficient manoeuvre decision. Here, CNN helps extract feature space, and two fully connected network generates longitudinal and lateral manoeuvre decisions. Performance evaluation of DST-CAN has been carried out using two publicly available NGSIM US-101 and I-80 highway datasets. A traffic rule is defined to generate ground truths for these datasets in addition to human decisions. Two DST-CAN models are trained using imitation learning with human driving decisions from actual traffic data and rule-based ground truth decisions. The performances of the DST-CAN models are compared with the state-of-the-art Convolutional Social-LSTM (CS-LSTM) models for manoeuvre prediction. The results clearly indicate that the context-aware maps help DST-CAN to predict the decision accurately over CS-LSTM. Further, an ablation study has been carried out to understand the effect of prediction horizons of performance and a robustness study to understand the near collision scenarios over actual traffic observations. The context-aware map with a 3 second prediction horizon is robust against near collision.

In this article, a priority-based dynamic resource Allocation with decentralized Multitask assignment (P-DREAM) approach is presented to protect a territory from highly manoeuvring intruders. In the first part, static optimization problems are formulated to compute the following parameters of the perimeter defense problem; the number of reserve stations, their locations, the priority region, the monitoring region, and the minimum number of defenders required for the monitoring purpose. The concept of a prioritized intruder is proposed here to identify and handle those critical intruders (computed based on the velocity ratio and location) to be tackled on a priority basis. The computed priority region helps to assign reserve defenders sufficiently earlier such that they can neutralize the prioritized intruders. The monitoring region defines the minimum region to be monitored and is sufficient enough to handle the intruders. In the second part, the earlier developed DREAM approach is modified to incorporate the priority of an intruder. The proposed P-DREAM approach assigns the defenders to the prioritized intruders as the first task. A convex territory protection problem is simulated to illustrate the P-DREAM approach. It involves the computation of static parameters and solving the prioritized task assignments with dynamic resource allocation. Monte-Carlo results were conducted to verify the performance of P-DREAM, and the results clearly show that the P-DREAM approach can protect the territory with consistent performance against highly manoeuvring intruders.

This paper presents a Dynamic Internal Predictive Power Scheduling (DIPPS) approach for optimizing power management in microgrids, particularly focusingon external power exchanges among diverse prosumers. DIPPS utilizes a dynamic objective function with a time-varying binary parameter to control the timing of power transfers to the external grid, facilitated by efficient usage of energy storage for surplus renewable power. The microgrid power scheduling problem is modeled as a mixed-integer nonlinear programmig (MINLP-PS) and subsequently transformed into a mixed-integer linear programming (MILP-PS) optimization through McCormick's relaxation to reduce the computational complexity. A predictive window with 6 data points is solved at an average of 0.92s, a 97.6% improvement over the 38.27s required for the MINLP-PS formulation, implying the numerical feasibility of the DIPPS approach for real-time implementation. Finally, the approach is validated against a static objective using real-world load data across three case studies with different time-varying parameters, demonstrationg the ability of DIPPS to optimize power exchanges and efficiently utilize distributed resources whie shifting the eexternal power transfers to specified time durations.

In this paper, a Priority-based Dynamic REsource Allocation with decentralized Multi-task assignment (P-DREAM) approach is presented to protect a territory from highly manoeuvring intruders. In the first part, static optimization problems are formulated to compute the following parameters of the perimeter defense problem; the number of reserve stations, their locations, the priority region, the monitoring region, and the minimum number of defenders required for the monitoring purpose. The concept of a prioritized intruder is proposed here to identify and handle those critical intruders (computed based on the velocity ratio and location) to be tackled on a priority basis. The computed priority region helps to assign reserve defenders sufficiently earlier such that they can neutralize the prioritized intruders. The monitoring region defines the minimum region to be monitored and is sufficient enough to handle the intruders. In the second part, the earlier developed DREAM approach is modified to incorporate the priority of an intruder. The proposed P-DREAM approach assigns the defenders to the prioritized intruders as the first task. A convex territory protection problem is simulated to illustrate the P-DREAM approach. It involves the computation of static parameters and solving the prioritized task assignments with dynamic resource allocation. Monte-Carlo results were conducted to verify the performance of P-DREAM, and the results clearly show that the P-DREAM approach can protect the territory with consistent performance against highly manoeuvring intruders.

Flight envelope protection for a high-performance aircraft poses a challenge to the designers and involves a time-consuming procedure to verify the provided protection. This paper presents a design approach to protect the aircraft from departure by a command path limiter for a rate command attitude hold controller in both the pitch and roll axes. In this approach, the maximum and minimum rates are scheduled as a function of the dynamic pressure on the basis of the open loop aircraft capabilities. This is then augmented with a novel angle of attack protection that comes into play only when the pilot inputs cause the aircraft to exceed the incidence on the positive or negative side (one sided protection), while maintaining the rate command attitude hold behavior within the normal operational bounds of angle of attack. Traditional methods of piloted simulation with a representative cohort of pilots can be time consuming to set up and may not give sufficient confidence whether a departure protection scheme is effective. To address this, a unique multi-modal search using genetic algorithm is developed to verify that this command path protection is able to achieve carefree maneuvering of a fighter aircraft in its entire flight envelope. The sequence of rapid pilot control inputs is coded into a chromosome. The multi-modal genetic algorithm then uses operators like cross-over and mutation on a starting population of chromosomes to evolve new inputs sequences which are then run to obtain the aircraft response. The cost function of the genetic algorithm which is constructed from the aircraft time response is designed to favor the search for multiple maxima which drive the aircraft to departure. The open domain ADMIRE model has been used to demonstrate the approach. Results indicate that the command path design proposed in this paper can be used to protect against departure and the novel multi-modal genetic algorithm helps to verify the departure protection.