Abstract: Conventionally, activity detection has been through one mode i.e., smart watch. Though it works in reasonable cases, there are chances of false positives considerably. Other approaches include surveillance which limits itself to object detection. Embodiments of present disclosure provide systems and methods for detecting activities performed by user from data captured from multiple sensors. A first input (FI) comprising accelerometer data, heart rate and gyroscope data and second input (SI) comprising video data are obtained. Features are extracted from FI and pre-processed for a first activity (FA) detection using activity prediction model. Frames from SI are processed for creating bounding box of user and resized thereof to extract pose coordinates vector. Distance between vector of pose coordinates and training vectors of pose coordinates stored in the system is computed and a second activity (SA) is detected accordingly. Both the FA and SA are validated for determining true and/or false positive.
Abstract: Automated software testing have become very crucial in this digital stage. It consumes a lot of manual effort and time of a tester. This disclosure relates generally to a method and system for automated generation of test scenarios and automation scripts. The system is configured to record details of the application at micro level (page level), index based on the page navigation. The system is further configured to create a mind map or a tree using a traverse algorithm. This creates the necessary test scenarios based on flows or page navigation. At the same time, the system captures all the underlying screen properties and labels. Connect the field type to the action library (built in). Also ensure dataset is created while capturing the recording. Further utilizing the test scenarios, data set and the screen properties to generate the automation script automatically.
Abstract: This disclosure relates generally to method and system for an adaptive filter based learning model for time series sensor signal classification on edge devices. The adaptive filter based learning model for time series sensor signal classification enables automated-computationally lightweight learning (significant reduction in computational resources) and inferring/classification in real-time or near-real-time on CPU/memory/battery life constrained edge devices. The disclosed techniques for time series sensor signal classification on edge devices characterizes the intrinsic signal processing properties of the input time series sensor signals using linear adaptive filtering and derivative spectrum to efficiently construct the adaptive filter based learning model based on standard classification algorithms for time series sensor signal classification.
Abstract: Causality is a crucial paradigm in several domains where observational data is available. Primary goal of Causal Inference (CI) is to uncover cause-effect relationship between entities. Conventional methods face challenges in providing an accurate CI framework due to cofounding and selection bias in multiple treatment scenario. The present disclosure computes a Propensity Score (PS) from a received CI data for the plurality of subjects under test for a treatment. A Generalized Propensity Score (GPS) is computed for a plurality of treatments corresponding to the plurality of subjects by using the PS. Further, a plurality of task batches are created using the GPS and given as input to the DNN for training. Errors in factual data and in balancing representation of the DNN are rectified using a novel loss function. The trained DNN is further used for predicting the counter factual treatment response corresponding to the factual treatment data.
Abstract: Systems and methods for assessing quality of input text using recurrent neural networks is disclosed. The system obtains input text from user and performs a comparison of each word from input text with words from dictionary (or trained data) to determine a closest recommended word for each word in the input text. The input text is further analyzed to determine context of each word based on at least a portion of input text, and based on determined context, at least one of correct sentences, incorrect sentences, and/or complex sentences are determined from the input text. Each word is converted to a vector based on concept(s) by comparing each word across sentences of input text to generate vectors set, and quality of the input text is assessed based on vectors set, the comparison, determined context and at least one of correct sentences, incorrect sentences, complex sentences, or combinations thereof.
Abstract: Sepsis is one of the most prevalent causes of mortality in Intensive Care Units (ICUs) and delayed treatment is associated with increase in death and financial burden. There is no single laboratory test or clinical sign that by itself can be considered diagnostic of sepsis. The present disclosure provides discriminating domain specific continuous and categorical features that can reliably classify a subject being monitored into a sepsis class or a normal class. A combination of physiological parameters, laboratory parameters and demographic details are used to extract the discriminating features. Even though the parameters may be sporadic in nature, the systems and methods of the present disclosure make use of a sliding time window to generate continuous features that capture the trend in the sporadic data; and a binning approach to generate categorical features to discriminate deviation from the normal class and facilitate timely treatment.
Abstract: Conventional hierarchical time-series clustering is highly time consuming process as time-series are characteristically lengthy. Moreover, finding right similarity measure providing best possible hierarchical cluster is critical to derive accurate inferences from the hierarchical clusters. Method and system for Auto Encoded Compact Sequences (AECS) based hierarchical time-series clustering that enables compact latent representation of time-series using an undercomplete multilayered Seq2Seq LSTM auto encoder followed by generating of HCs using multiple similarity measures is disclosed. Further, provided is a mechanism to select the best HC among the multiple HCs on-the-fly, based on an internal clustering performance measure of Modified Hubert statistic ?. Thus, the method provides time efficient and low computational cost approach for hierarchical clustering for both on univariate and multivariate time-series.
March 22, 2021
October 14, 2021
Tata Consultancy Services Limited
Soma Bandyopadhyay, Anish Datta, Arpan Pal
Abstract: Systems and methods for generating control system solutions for robotics environments is provided. The traditional systems and methods provide robotics solutions but specialized to only a particular robotic application, domain, and selected structure.
Abstract: Augmented Reality is having numerous applications in the field of beverage industry. For example, in preparing cocktails. Conventional methods fail to provide an interactive AR based cocktail preparation, enabling the user to prepare cocktail. The present disclosure analyzes an image of a shelf to obtain a plurality of beverage types. A plurality of potential cocktails are generated based on the analysis and displayed to the user using 3D Augmented Reality (AR). A procedure for preparing a cocktail, requested by the user, is displayed near a located glassware kept by the user using 3D AR. A volume of the glassware is computed and a quantity of each beverage to be added is computed based on the volume. Further, a quantity of each beverage added by the user is dynamically measured by using 3D AR techniques and recommendations are provided to the user if there is any deviation.
February 12, 2021
Date of Patent:
October 12, 2021
Tata Consultancy Services Limited
Robin Tommy Kulangara Muriyil, Reshmi Ravindranathan, Zakhir Sidickk, Anantha Lakshmeeswari Chandarla
Abstract: Traditional cognitive load estimation techniques rely on raw pupil size alone which is often prone to confound with changes in illumination, errors associated with sensor devices and irregular oscillations of pupil under constant light conditions. Estimation of cognitive load finds application in many domains including optimum work allocation, assessing a work environment and medical diagnosis. The present disclosure employs frequency domain analysis of pupil size variations to estimate load imposed by a cognitive task. A cognitive load metric based on power and frequency relations at mean frequency of the variation in pupil size addresses cognitive load estimation based on pupil dilation, wherein the pupil dilation is captured by employing low cost non-intrusive nearables.
Abstract: A data driven approach for fault detection in robotic actuation is disclosed. Here, a set of robotic tasks are received and analyzed by a Deep Learning (DL) analytics. The DL analytics includes a stateful (Long Short Term Memory) LSTM. Initially, the stateful LSTM is trained to match a set of activities associated with the robots based on a set of tasks gathered from the robots in a multi robot environment. Here, the stateful LSTM utilizes a master slave framework based load distribution technique and a probabilistic trellis approach to predict a next activity associated with the robot with minimum latency and increased accuracy. Further, the predicted next activity is compared with an actual activity of the robot to identify any faults associated robotic actuation.
Abstract: In the field of Internet of Things understanding need of applications and translating them to network parameters and protocol parameters is a major challenge. This disclosure addresses problem of enabling network services by cognitive sense-analyze-decide-respond framework. A processor implemented method is provided for enabling network aware applications and applications aware networks by a sense analyze decide respond (SADR) framework.
Abstract: Conventionally, loading and unloading pallets occurred in a factory like environment, wherein traditional systems were used for stacking pallets vertically for retrieval thereof. However, these pallets arrive from a roller conveyer and such set-ups are purely based on the concept of storage purpose and lack in palletizing and depalletizing in effective manner. Embodiments of the present disclosure provide pallet loading and unloading apparatus for automated objects manipulation, wherein pallet(s) from a forklift jack arrives on a linear slider assembly for movement of pallet from loading area to manipulating area using two double-sided cylinders associated thereof. Rollers comprised in apparatus provide flexibility to linear slider to slide freely on top. Manipulator performs palletizing and depalletizing as applicable.
Abstract: This disclosure relates generally to a system and method for auto-generation of test specifications from internet of things (IoT) solution specifications of IoT-enabled components of an IoT network. Testing is the complementary and most important part of any IoT network. Herein, a domain specific language (DSL) is used to specify capability of IoT enabled components. IoT solution specifications are captured from capabilities of IoT enabled components using a predefined activity DSL. A flow of activity is captured to assert transitions among one or more activities based on guard conditions. The flow of activity is analyzed to generate test specifications automatically using a Test Specification DSL based on the asserted transitions. The test specifications are implemented automatically in a predefined target language corresponding to the IoT enabled components.
September 15, 2020
September 30, 2021
Tata Consultancy Services Limited
Barnali Basak, Subhrojyoti Roy Chaudhuri
Abstract: Existing wearable device-based approaches to capture tremor signal have accuracy limitations due to usage of accelerometer sensor with inherent noisy nature. The method and system disclosed herein taps characteristics of the PPG sensor of being sensitive to the motion artefact, as an advantage, to capture tremor signal present in the PPG sensor. The method disclosed herein describes an approach to extract tremor signal of interest from the PPG signal by performing a Singular Spectrum Analysis (SSA) followed by spectrum density estimation. The SSA comprises performing embedding on the acquired PPG signal, performing Principal Component Analysis (PCA) on the embedded signal and reconstructing the rest tremor signal from the significant principal components identified post the PCA. Further, the spectrum density estimation detects a dominant frequency present in the principal components, which is the dominant frequency associated with the rest tremor.
Abstract: This disclosure relates generally to system and method for optimization of industrial processes, for example a tundish process. Typically geometries for industrial processes are simulated in a numerical analysis model such as a CFD. In order to simulate a physical phenomenon (such as tundish process) numerically, the domain thereof is discretized in order to convert the differential equations to be solved in the domain into linear equations. The accuracy of a CFD solution is dependent on a mesh of the domain, which in turn depends on a geometry thereof. For setting up an optimization task, the disclosed method provides first a CFD friendly base geometry, so that a faulty geometry can be detected before forming the complete geometry.
Abstract: Embodiments herein provide a method and system for continuously validating a user during an established authenticated session using Photoplethysmogram (PPG) and accelerometer data. State of the art approaches are mostly based on feature extraction and ML modelling for PPG based continuous session validation, while a template based approach in the art follows a complicated approach. The method disclosed herein utilizes less computation intensive template based approach to continuously validate the user across the session. The method comprises preprocessing a PPG data or PPG signal acquired from a wearable device worn by the user to identify segments of negligible motion. A first segment, after authentication using conventional authentication mechanism, serves as the initial reference. The chosen segments are then tested one by one with respect to the reference. If the templates in a segment match those of the reference, it is updated as the new reference, else a re-authentication is triggered.
Abstract: This disclosure relates generally to a system and method to estimate an operational risk associated with one or more failures in at least one unit of a process plant. There is a continuous stream of operational data of several variables such as temperature, pressure, etc. Detections are defined in terms of acceptable/unacceptable ranges of parameters over a finite period and operating load of the unit. Often, these predefined parameters must be within a specified range based on operating condition of the process plant and when the measured parameters go beyond, a failure is detected. A risk priority number is estimated from number of occurrences of failure mode, average percentage change from dynamic limits with severity and degree of correlation with detectability from operational data and dynamic limits. Herein, operational risk associated with failure modes can be calculated and updated from time to time automatically from the stream of operational data.
Abstract: The disclosure generally relates to methods and systems for identifying presence of abnormal heart sounds from heart sound signals of a subject being monitored. Conventional Artificial intelligence (AI) based abnormal heart sounds detection models with supervised learning requires a substantial amount of accurate training datasets covering all heart disease types for the training, which is quiet challenging. The present methods and systems solve the problem solves the problem of identifying presence of the abnormal heart sounds using an efficient semi-supervised learning model. The semi-supervised learning model is generated based on probability distribution of spectrographic properties obtained from heart sound signals of healthy subjects. A Kullback-Leibler (KL) divergence between a predefined Gaussian distribution and an encoded probability distribution of the semi-supervised learning model is determined as an anomaly score for identifying the abnormal heart sounds.
Abstract: The disclosure generally relates to determining a breathing rate and a heart rate from cardiopulmonary signal. Conventional systems use additional hardware to improve signal quality or different signal processing techniques to calculate the heart rate from the cardiopulmonary signal. However accurately determining the heart rate is always a continuous area of an improvement. The present methods and systems solve the problem of determining the heart rate accurately, from the cardiopulmonary signals, by determining the signal quality of the cardiopulmonary signal and the signal associated with the heart rate, comprised in the cardiopulmonary signal. A signal processing technique that best performs, out of a set of signal processing techniques is identified based on the signal quality to determine the heart rate. A long-term and an effective health monitoring of healthy as well as patient and infant subjects is achieved by the present disclosure.