arXiv subjects

Suresh Sundaram

Publications and source records attributed to Suresh Sundaram.

At least 73 records · Page 4Linked to original sources

arXiv2022

Robust EMRAN-aided Coupled Controller for Autonomous Vehicles

This paper presents a coupled, neural network-aided longitudinal cruise and lateral path-tracking controller for an autonomous vehicle with model uncertainties and experiencing unknown external disturbances. Using a feedback error learning mechanism, an inverse vehicle dynamics learning scheme utilizing an adaptive Radial Basis Function (RBF) neural network, referred to as the Extended Minimal Resource Allocating Network (EMRAN) is employed. EMRAN uses an extended Kalman filter for online learning and weight updates, and also incorporates a growing/pruning strategy for maintaining a compact network for easier real-time implementation. The online learning algorithm handles the parametric uncertainties and eliminates the effect of unknown disturbances on the road. Combined with a self-regulating learning scheme for improving generalization performance, the proposed EMRAN-aided control architecture aids a basic PID cruise and Stanley path-tracking controllers in a coupled form. Its performance and robustness to various disturbances and uncertainties are compared with the conventional PID and Stanley controllers, along with a comparison with a fuzzy-based PID controller and an active disturbance rejection control (ADRC) scheme. Simulation results are presented for both slow and high speed scenarios. The root mean square (RMS) and maximum tracking errors clearly indicate the effectiveness of the proposed control scheme in achieving better tracking performance in autonomous vehicles under unknown environments.

eess.SY↗

arXiv2021

Fast Obstacle Avoidance Motion in SmallQuadcopter operation in a Cluttered Environment

The autonomous operation of small quadcopters moving at high speed in an unknown cluttered environment is a challenging task. Current works in the literature formulate it as a Sense-And-Avoid (SAA) problem and address it by either developing new sensing capabilities or small form-factor processors. However, the SAA, with the high-speed operation, remains an open problem. The significant complexity arises due to the computational latency, which is critical for fast-moving quadcopters. In this paper, a novel Fast Obstacle Avoidance Motion (FOAM) algorithm is proposed to perform SAA operations. FOAM is a low-latency perception-based algorithm that uses multi-sensor fusion of a monocular camera and a 2-D LIDAR. A 2-D probabilistic occupancy map of the sensing region is generated to estimate a free space for avoiding obstacles. Also, a local planner is used to navigate the high-speed quadcopter towards a given target location while avoiding obstacles. The performance evaluation of FOAM is evaluated in simulated environments in Gazebo and AIRSIM. Real-time implementation of the same has been presented in outdoor environments using a custom-designed quadcopter operating at a speed of $4.5$ m/s. The FOAM algorithm is implemented on a low-cost computing device to demonstrate its efficacy. The results indicate that FOAM enables a small quadcopter to operate at high speed in a cluttered environment efficiently.

cs.RO↗

arXiv2021

Confidence Conditioned Knowledge Distillation

In this paper, a novel confidence conditioned knowledge distillation (CCKD) scheme for transferring the knowledge from a teacher model to a student model is proposed. Existing state-of-the-art methods employ fixed loss functions for this purpose and ignore the different levels of information that need to be transferred for different samples. In addition to that, these methods are also inefficient in terms of data usage. CCKD addresses these issues by leveraging the confidence assigned by the teacher model to the correct class to devise sample-specific loss functions (CCKD-L formulation) and targets (CCKD-T formulation). Further, CCKD improves the data efficiency by employing self-regulation to stop those samples from participating in the distillation process on which the student model learns faster. Empirical evaluations on several benchmark datasets show that CCKD methods achieve at least as much generalization performance levels as other state-of-the-art methods while being data efficient in the process. Student models trained through CCKD methods do not retain most of the misclassifications commited by the teacher model on the training set. Distillation through CCKD methods improves the resilience of the student models against adversarial attacks compared to the conventional KD method. Experiments show at least 3% increase in performance against adversarial attacks for the MNIST and the Fashion MNIST datasets, and at least 6% increase for the CIFAR10 dataset.

cs.LG↗

arXiv2021

Generative Replay-based Continual Zero-Shot Learning

Zero-shot learning is a new paradigm to classify objects from classes that are not available at training time. Zero-shot learning (ZSL) methods have attracted considerable attention in recent years because of their ability to classify unseen/novel class examples. Most of the existing approaches on ZSL works when all the samples from seen classes are available to train the model, which does not suit real life. In this paper, we tackle this hindrance by developing a generative replay-based continual ZSL (GRCZSL). The proposed method endows traditional ZSL to learn from streaming data and acquire new knowledge without forgetting the previous tasks' gained experience. We handle catastrophic forgetting in GRCZSL by replaying the synthetic samples of seen classes, which have appeared in the earlier tasks. These synthetic samples are synthesized using the trained conditional variational autoencoder (VAE) over the immediate past task. Moreover, we only require the current and immediate previous VAE at any time for training and testing. The proposed GRZSL method is developed for a single-head setting of continual learning, simulating a real-world problem setting. In this setting, task identity is given during training but unavailable during testing. GRCZSL performance is evaluated on five benchmark datasets for the generalized setup of ZSL with fixed and dynamic (incremental class) settings of continual learning. The existing class setting presented recently in the literature is not suitable for a class-incremental setting. Therefore, this paper proposes a new setting to address this issue. Experimental results show that the proposed method significantly outperforms the baseline and the state-of-the-art method and makes it more suitable for real-world applications.

cs.CV↗

arXiv2021

Self Regulated Learning Mechanism for Data Efficient Knowledge Distillation

Existing methods for distillation do not efficiently utilize the training data. This work presents a novel approach to perform distillation using only a subset of the training data, making it more data-efficient. For this purpose, the training of the teacher model is modified to include self-regulation wherein a sample in the training set is used for updating model parameters in the backward pass either if it is misclassified or the model is not confident enough in its prediction. This modification restricts the participation of samples, unlike the conventional training method. The number of times a sample participates in the self-regulated training process is a measure of its significance towards the model's knowledge. The significance values are used to weigh the losses incurred on the corresponding samples in the distillation process. This method is named significance-based distillation. Two other methods are proposed for comparison where the student model learns by distillation and incorporating self-regulation as the teacher model, either utilizing the significance information computed during the teacher's training or not. These methods are named hybrid and regulated distillations, respectively. Experiments on benchmark datasets show that the proposed methods achieve similar performance as other state-of-the-art methods for knowledge distillation while utilizing a significantly less number of samples.

cs.LG↗

arXiv2021

Scaffolding Reflection in Reinforcement Learning Framework for Confinement Escape Problem

In this paper, a novel Scaffolding Reflection in Reinforcement Learning (SR2L) is proposed for solving the confinement escape problem (CEP). In CEP, an evader's objective is to attempt escaping a confinement region patrolled by multiple pursuers. Meanwhile, the pursuers aim to reach and capture the evader. The inverse solution for pursuers to try and capture has been extensively studied in the literature. However, the problem of evaders escaping from the region is still an open issue. The SR2L employs an actor-critic framework to enable the evader to escape the confinement region. A time-varying state representation and reward function have been developed for proper convergence. The formulation uses the sensor information about the observable environment and prior knowledge of the confinement boundary. The conventional Independent Actor-Critic (IAC) method fails to converge due to sparseness in the reward. The effect becomes evident when operating in such a dynamic environment with a large area. In SR2L, along with the developed reward function, we use the scaffolding reflection method to improve the convergence significantly while increasing its efficiency. In SR2L, a motion planner is used as a scaffold for the actor-critic network to observe, compare and learn the action-reward pair. It enables the evader to achieve the required objective while using lesser resources and time. Convergence studies show that SR2L learns faster and converges to higher rewards as compared to IAC. Extensive Monte-Carlo simulations show that a SR2L consistently outperforms conventional IAC and the motion planner itself as the baselines.

cs.RO↗

arXiv2021

Online Lifelong Generalized Zero-Shot Learning

Methods proposed in the literature for zero-shot learning (ZSL) are typically suitable for offline learning and cannot continually learn from sequential streaming data. The sequential data comes in the form of tasks during training. Recently, a few attempts have been made to handle this issue and develop continual ZSL (CZSL) methods. However, these CZSL methods require clear task-boundary information between the tasks during training, which is not practically possible. This paper proposes a task-free (i.e., task-agnostic) CZSL method, which does not require any task information during continual learning. The proposed task-free CZSL method employs a variational autoencoder (VAE) for performing ZSL. To develop the CZSL method, we combine the concept of experience replay with knowledge distillation and regularization. Here, knowledge distillation is performed using the training sample's dark knowledge, which essentially helps overcome the catastrophic forgetting issue. Further, it is enabled for task-free learning using short-term memory. Finally, a classifier is trained on the synthetic features generated at the latent space of the VAE. Moreover, the experiments are conducted in a challenging and practical ZSL setup, i.e., generalized ZSL (GZSL). These experiments are conducted for two kinds of single-head continual learning settings: (i) mild setting-: task-boundary is known only during training but not during testing; (ii) strict setting-: task-boundary is not known at training, as well as testing. Experimental results on five benchmark datasets exhibit the validity of the approach for CZSL.

cs.CV↗

arXiv2021

Spatio-Temporal Look-Ahead Trajectory Prediction using Memory Neural Network

Prognostication of vehicle trajectories in unknown environments is intrinsically a challenging and difficult problem to solve. The behavior of such vehicles is highly influenced by surrounding traffic, road conditions, and rogue participants present in the environment. Moreover, the presence of pedestrians, traffic lights, stop signs, etc., makes it much harder to infer the behavior of various traffic agents. This paper attempts to solve the problem of Spatio-temporal look-ahead trajectory prediction using a novel recurrent neural network called the Memory Neuron Network. The Memory Neuron Network (MNN) attempts to capture the input-output relationship between the past positions and the future positions of the traffic agents. The proposed model is computationally less intensive and has a simple architecture as compared to other deep learning models that utilize LSTMs and GRUs. It is then evaluated on the publicly available NGSIM dataset and its performance is compared with several state-of-art algorithms. Additionally, the performance is also evaluated on a custom synthetic dataset generated from the CARLA simulator. It is seen that the proposed model outperforms the existing state-of-art algorithms. Finally, the model is integrated with the CARLA simulator to test its robustness in real-time traffic scenarios.

cs.RO↗

arXiv2021

Design and Integration of a Drone based Passive Manipulator for Capturing Flying Targets

In this paper, we present a novel passive single Degree-of-Freedom (DoF) manipulator design and its integration on an autonomous drone to capture a moving target. The end-effector is designed to be passive, to disengage the moving target from a flying UAV and capture it efficiently in the presence of disturbances, with minimal energy usage. It is also designed to handle target sway and the effect of downwash. The passive manipulator is integrated with the drone through a single Degree of Freedom (DoF) arm, and experiments are carried out in an outdoor environment. The rack-and-pinion mechanism incorporated for this manipulator ensures safety by extending the manipulator beyond the body of the drone to capture the target. The autonomous capturing experiments are conducted using a red ball hanging from a stationary drone and subsequently from a moving drone. The experiments show that the manipulator captures the target with a success rate of 70\% even under environmental/measurement uncertainties and errors.

cs.RO↗

arXiv2021

Design Iterations for Passive Aerial Manipulator

Grabbing a manoeuvring target using drones is a challenging problem. This paper presents the design, development, and prototyping of a novel aerial manipulator for target interception. It is a single Degree of Freedom (DoF) manipulator with passive basket-type end-effector. The proposed design is energy efficient, light weight and suitable for aerial grabbing applications. The detailed design of the proposed manipulation mechanism and a novel in-flight extending propeller guard, is reported in this paper.

cs.RO↗

arXiv2021

A Decentralized Multi-UAV Spatio-Temporal Multi-Task Allocation Approach for Perimeter Defense

This paper provides a new solution approach to a multi-player perimeter defense game, in which the intruders' team tries to enter the territory, and a team of defenders protects the territory by capturing intruders on the perimeter of the territory. The objective of the defenders is to detect and capture the intruders before the intruders enter the territory. Each defender independently senses the intruder and computes his trajectory to capture the assigned intruders in a cooperative fashion. The intruder is estimated to reach a specific location on the perimeter at a specific time. Each intruder is viewed as a spatio-temporal task, and the defenders are assigned to execute these spatio-temporal tasks. At any given time, the perimeter defense problem is converted into a Decentralized Multi-UAV Spatio-Temporal Multi-Task Allocation (DMUST-MTA) problem. The cost of executing a task for a trajectory is defined by a composite cost function of both the spatial and temporal components. In this paper, a decentralized consensus-based bundle algorithm has been modified to solve the spatio-temporal multi-task allocation problem, and the performance evaluation of the proposed approach is carried out based on Monte-Carlo simulations. The simulation results show the effectiveness of the proposed approach to solve the perimeter defense game under different scenarios. Performance comparison with a state-of-the-art centralized approach with full observability, clearly indicates that DMUST-MTA achieves similar performance in a decentralized way with partial observability conditions with a lesser computational time and easy scaling up.

eess.SY↗

arXiv2021

Generalized Continual Zero-Shot Learning

Recently, zero-shot learning (ZSL) emerged as an exciting topic and attracted a lot of attention. ZSL aims to classify unseen classes by transferring the knowledge from seen classes to unseen classes based on the class description. Despite showing promising performance, ZSL approaches assume that the training samples from all seen classes are available during the training, which is practically not feasible. To address this issue, we propose a more generalized and practical setup for ZSL, i.e., continual ZSL (CZSL), where classes arrive sequentially in the form of a task and it actively learns from the changing environment by leveraging the past experience. Further, to enhance the reliability, we develop CZSL for a single head continual learning setting where task identity is revealed during the training process but not during the testing. To avoid catastrophic forgetting and intransigence, we use knowledge distillation and storing and replay the few samples from previous tasks using a small episodic memory. We develop baselines and evaluate generalized CZSL on five ZSL benchmark datasets for two different settings of continual learning: with and without class incremental. Moreover, CZSL is developed for two types of variational autoencoders, which generates two types of features for classification: (i) generated features at output space and (ii) generated discriminative features at the latent space. The experimental results clearly indicate the single head CZSL is more generalizable and suitable for practical applications.

cs.CV↗

arXiv2020

Gap Reduced Minimum Error Robust Simultaneous Estimation For Unstable Nano Air Vehicle

This paper proposes a novel Gap Reduced Minimum Error Robust Simultaneous (GRMERS) estimator for resource-constrained Nano Aerial Vehicle (NAV) that enables a single estimator to provide simultaneous and robust estimation for a given N unstable and uncertain NAV plant models. The estimated full state feedback enables a stable flight for NAV. The GRMERS estimator is implemented utilizing a Minimum Error Robust Simultaneous (MERS) estimator and Gap Reducing (GR) compensators. The MERS estimator provides robust simultaneous estimation with minimal largest worst-case estimation error even in the presence of a bounded energy exogenous disturbance signal. The GR compensators reduce the gap between the graphs of N linear plant models to decrease the estimation error generated by the MERS estimator. A sufficient condition for the existence of a simultaneous estimator is established using LMIs and robust estimation theory. Further, MERS estimator and GR compensator design are formulated as non-convex tractable optimization problems and are solved using the population-based genetic algorithms. The performance of the GRMERS estimator consisting of MERS estimator and GR compensators from the population-based genetic algorithms is validated through simulation studies. The study results indicate that a single GRMERS estimator can produce state estimates with reduced errors for all flight conditions. The results indicate that the single GRMERS estimator is robust than the individually designed H inifinity filters.

eess.SY↗

arXiv2020

Meta-Cognition-Based Simple And Effective Approach To Object Detection

Recently, many researchers have attempted to improve deep learning-based object detection models, both in terms of accuracy and operational speeds. However, frequently, there is a trade-off between speed and accuracy of such models, which encumbers their use in practical applications such as autonomous navigation. In this paper, we explore a meta-cognitive learning strategy for object detection to improve generalization ability while at the same time maintaining detection speed. The meta-cognitive method selectively samples the object instances in the training dataset to reduce overfitting. We use YOLO v3 Tiny as a base model for the work and evaluate the performance using the MS COCO dataset. The experimental results indicate an improvement in absolute precision of 2.6% (minimum), and 4.4% (maximum), with no overhead to inference time.

cs.CV↗

arXiv2020

Restricted Airspace Protection using Multi-UAV Spatio-TemporalMulti-Task Allocation

This paper addresses the problem of restricted airspace protection from invaders using the cooperative multi-UAV system. The objective is to detect and capture the invaders cooperatively by a team of homogeneous UAVs (called evaders)before invaders enter the restricted airspace. The problem of restricted airspace protection problem is formulated as a Multi-UAV Spatio-Temporal Multi-Task Allocation problem and is referred as MUST-MTA. The MUST-MTA problem is solved using a modified consensus-based bundled auction method. Here, the spatial and time constraints are handled by combining both spatial and temporal loss component. The solution identifies the sequence of spatial locations to be reached by the evader at specific time instants to neutralize the invaders. The performance of MUST-MTA with the consensus approach is evaluated in a simulated environment. The Monte-Carlo simulation results clearly indicate the efficacy of the proposed approach in restricted airspace protection against intruders

eess.SY↗

arXiv2020

Highly Ordered Boron Nitride/Epigraphene Epitaxial Films on Silicon Carbide by Lateral Epitaxial Deposition

Realizing high-performance nanoelectronics requires control of materials at the nanoscale. Methods to produce high quality epitaxial graphene (EG) nanostructures on silicon carbide are known. The next step is to grow Van der Waals semiconductors on top of EG nanostructures. Hexagonal boron nitride (h-BN) is a wide bandgap semiconductor with a honeycomb lattice structure that matches that of graphene, making it ideally suited for graphene-based nanoelectronics. Here, we describe the preparation and characterization of multilayer h-BN grown epitaxially on EG using a migration-enhanced metalorganic vapor phase epitaxy process. As a result of the lateral epitaxial deposition (LED) mechanism, the grown h-BN/EG heterostructures have highly ordered epitaxial interfaces, as desired in order to preserve the transport properties of pristine graphene. Atomic scale structural and energetic details of the observed row-by-row, growth mechanism of the 2D epitaxial h-BN film are analyzed through first-principles simulations, demonstrating one-dimensional nucleation-free-energy-barrierless growth. This industrially relevant LED process can be applied to a wide variety of van der Waals materials.

cond-mat.mtrl-sci↗

arXiv2020

CGAP2: Context and gap aware predictive pose framework for early detection of gestures

With a growing interest in autonomous vehicles' operation, there is an equally increasing need for efficient anticipatory gesture recognition systems for human-vehicle interaction. Existing gesture-recognition algorithms have been primarily restricted to historical data. In this paper, we propose a novel context and gap aware pose prediction framework(CGAP2), which predicts future pose data for anticipatory recognition of gestures in an online fashion. CGAP2 implements an encoder-decoder architecture paired with a pose prediction module to anticipate future frames followed by a shallow classifier. CGAP2 pose prediction module uses 3D convolutional layers and depends on the number of pose frames supplied, the time difference between each pose frame, and the number of predicted pose frames. The performance of CGAP2 is evaluated on the Human3.6M dataset with the MPJPE metric. For pose prediction of 15 frames in advance, an error of 79.0mm is achieved. The pose prediction module consists of only 26M parameters and can run at 50 FPS on the NVidia RTX Titan. Furthermore, the ablation study indicates supplying higher context information to the pose prediction module can be detrimental for anticipatory recognition. CGAP2 has a 1-second time advantage compared to other gesture recognition systems, which can be crucial for autonomous vehicles.

cs.CV↗

arXiv2020

Context-Aware Deep Q-Network for Decentralized Cooperative Reconnaissance by a Robotic Swarm

One of the crucial problems in robotic swarm-based operation is to search and neutralize heterogeneous targets in an unknown and uncertain environment, without any communication within the swarm. Here, some targets can be neutralized by a single robot, while others need multiple robots in a particular sequence to neutralize them. The complexity in the problem arises due to the scalability and information uncertainty, which restricts the robot's awareness of the swarm and the target distribution. In this paper, this problem is addressed by proposing a novel Context-Aware Deep Q-Network (CA-DQN) framework to obtain communication free cooperation between the robots in the swarm. Each robot maintains an adaptive grid representation of the vicinity with the context information embedded into it to keep the swarm intact while searching and neutralizing the targets. The problem formulation uses a reinforcement learning framework where two Deep Q-Networks (DQNs) handle 'conflict' and 'conflict-free' scenarios separately. The self-play-in-based approach is used to determine the optimal policy for the DQNs. Monte-Carlo simulations and comparison studies with a state-of-the-art coalition formation algorithm are performed to verify the performance of CA-DQN with varying environmental parameters. The results show that the approach is invariant to the number of detected targets and the number of robots in the swarm. The paper also presents the real-time implementation of CA-DQN for different scenarios using ground robots in a laboratory environment to demonstrate the working of CA-DQN with low-power computing devices.

cs.MA↗