Author :-Tanya juneja
IOT (Internet of Things)-This term was first coined in 1999 by Kevin Ashton of Proctor and gamble. It is network of physical object that has sensor in it or any software or other technology which enable us connecting and exchanging data with other system with the help of internet. These kinds of devices generate massive data. This data can be used for analysis purpose of further patterns, predictions and assessments.
As the evolution begin merging of various technology together like real time analytics, machine learning , wireless network system, embedded system and etc. A new mechanism is required to manage the values of IoT generated data. The most suitable computational paradigm which help is Machine Learning (ML). Machine learning provides embedded intelligence to IoT devices and infer the knowledge from device generated data.ML can be used in classification, regression and density estimation. ML techniques and algorithm can be leveraged in IoT based applications such as fraud detection, malware detection and speech recognition and provide intelligent services.
Threat to IOT devices:-
IoT has great influence by creating a new dimension in the internet world. The main threat in IoT services and application are security and privacy. Security of IoT device includes architectural security, data security, communication security, malware analysis and so on. A cross layer design and optimized algorithm are required as the solution to security and privacy issues of IoT. IoT devices further may require a new breed of cryptographic and other algorithm to cope up with security and privacy. However, increase in number of IoT devices may increase the challenges in security mechanism. A holistic security and privacy approach is considered more than other existing security solution. This new approach helps in providing new intelligent, robust, evolutionary and scalable mechanism to control the challenges and insecurity in IoT applications.
ML is one of those intelligent methods that provide the optimal solution through learning using past experience or example data. ML uses mathematical techniques to build models of behaviours. ML can also enable the smart devices to learn without use of explicitly programming. ML has multidisciplinary nature that may include artificial Intelligence, optimization theory and cognitive science. This feature allows ML to be useful where human expertise cannot be used such as robotics and speech recognition and real life problem. It provides the solution to IoT where solution to a specific problem changes in time. Though ML technique is very reliable but sometimes it produces false positive and true negative results so it requires proper guidance and modification. An advance version of ML that is Deep Learning (DL) helps to overcome this issue and determines the accuracy of prediction by itself.
IoT Machine Learning Application
These are important IoT and Machine Learning Use Cases:
Value saving in industrial programs
Predictive talents are used in a mechanical putting. By getting information from unique sensors in or on machines, machine calculations can help us understand what’s common for the machine and in a while let us discover when something unexpected event begins to arise or happen. A system needs protection is a crucial aim .Businesses usually make use of machine learning and are expects over 90% of accuracy of machines which will further need renovation, meaning large fee cuttings.
FIG:1
Shaping Experiences to Individuals
As we all know Amazon, Netflix and other such websites make use of machine learning figuring out how to absorb our pattern, behavior and give us a superior ordeal. They will give you suggestions regarding gadgets that you would like or giving proposals to films, tv serial, etc. So additionally, IoT machine learning would be able to understand pattern and behavior forming our situation to our own dispositions or ideas. One example is The Nest Thermostat is an incredible case, it utilizes machine identifying the way to soak up your dispositions for warming and cooling, making sure that the house is the best temperature whilst you go back home from work or while you stand up inside the morning.
ML and IoT interaction using python:-
There are various project idea by which would enable you to use IOT and ML together .You just have to load your device or censor data using panda library .Then your next step is to use data mining criteria and at last you have to build a model that is best suited for solving your problem. The problem taken here is human face detection using cameras.
Applications for face detection use algorithms that rely on identifying human faces in broader photos that may include environments, artifacts, and other sections of a person's physique. This work proposes a real-time identification system, based on modern image processing capabilities of open source API like OpenCV and due to the solution requirements, a study on the performance analysis of such solution compared to available commercial framework like SPID from NEC is intended. However, here, the study is available with the results of various experiments on the developed system. A systematic approach is followed to produce such outputs and have been measured using software codes. By using IP camera and a Raspberry Pi, the solution developed is simple in nature. This study relies on face detection and identification functionalities for human faces but not limited to live faces only but mix of faces from still images as well.
Stepwise:-
To create dataset for human face .
Firstly,face is detected. Face detection is a computer technology being used in a variety of applications that identifies human faces in digital images.
For this harcasscade algorithm is used. It is based on the Haar Wavelet technique to analyze pixels in the image into squares by function.
Haar Cascades use the Adaboost learning algorithm which selects a small number of important features from a large set to give an efficient result of classifiers.
This uses machine learning techniques to get a high degree of accuracy from what is called “training data”.
FIG:3
To store different dataset repeating the above process for various individual.
To perform face recognition.
Face recognition is a method of identifying or verifying the identity of an individual using their face. Face recognition systems can be used to identify people in photos, video, or in real-time.
For doing this we can use fisherface algorithm
Fisherfaces algorithm extracts principle components that separate one individual from another. So, now an individual's features can't dominate another person's features.
Fisherface method will be applied to generate feature vector of facial image data used by system and then to match vector of traits of training image with vector characteristic of test image using euclidean distance formula.
To perform classification using SVM(support vector machine and hog) for further classification.
HOG descriptor
The HOG descriptor models the local object texture by calculating the distribution of the local intensity gradients and the edge directions. It is formed by well-normalized local histograms of image gradient orientations in a dense grid. The algorithm starts with dividing the image into n small cells. For each cell, an m-bin HOG is built. Each bin in the HOG corresponds to an orientation spanning. The combination of n histograms forms the HOG descriptor, with the size of m∗ n bins. To calculate the HOG, the intensity gradients are calculated first in both the horizontal and vertical directions. Then, the magnitude and the orientation of the gradient is calculated. Each gradient has a vote in its bin, which is its magnitude.
Model Training and Classification
The detector is built by feeding the sets of positive and negative samples into an SVM. For each detector and parameter combination, a preliminary detector is trained. Then the false positives are picked out and used as the “hard samples” to augment the training set. The model is then retrained using this augmented set to produce the final detector. As claimed in [6], this retraining process significantly improves the performance
of the detector.
LIMITATION OF USING MACHINE LEARNING IN IoT
There are certain uncertainties associated with IoT data which require modification, However ML technologies are not efficient enough for certain failure or breakpoints. Therefore, a few limitations are also there while using ML in IoT.
Machine learning algorithms have issues regarding memory, and computational complexity.
ML techniques are only limited to low dimensional problem as they lack in scalability.
ML works on constant stream of data in real time so sometimes it is not suitable for smart IoT devices which work on real time data processing.
The predictive ability of an algorithm decreases with the increase in the dimensionality of data.
IoT network generates diversify data which differ in semantics and format, hence exhibit syntactic and semantic heterogeneity which raises a problematic issue to ML. In the case of ML statistical properties of entire dataset remain the same therefore, in real time applications where data from various sources have been different in formatting and representations causes problems for
ML algorithms as these algorithms are not efficient to work on semantic and syntactic diversified data. Merging of ML algorithm with existing streaming solutions enhances the overall complexity of an algorithm.
Futuristic Scope of IoT and Machine Learning
ML is the driving for artificial intelligence. The major advantages of using ML system are heuristic learning, decision tree for Administration purpose and data acquisition . All data science based applications like data mining, information retrieval system, search engine and big data analysis use ML algorithms. It also help to find applications in computer vision for object identification. IoT is the most used and recent application of ML. Many researchers presented in heir survey using IoT with ML in different applications and services. IoT in cyber security system using data mining techniques. Most of the proposed schemes focuses on technical area which are not on the user’s needs sometime. Though ML helps to reduce burden of user in their daily life by providing them technological advantages, yet raises some critical issues. In future ML with combination of IoT needs to work on several critical issues such as scalability, cost, battery of sensors, handling of multiple sensors, time elapse and many more.
References:-
https://en.wikipedia.org/wiki/Internet_of_things
https://data-flair.training/blogs/iot-and-machine-learning/
https://www.citlprojects.com/ieee-machine-learning-project-list/iot-ml-ai-deep-learning
Commenti