Nowadays, Wireless Sensor Network (WSN) is a modern technology with a wide range of applications and greatly attractive benefits, for example, self-governing, low expenditure on execution and data communication, long-term function, and unsupervised access to the network. The Internet of Things (IoT) is an attractive, exciting paradigm. By applying communication technologies in sensors and supervising features, WSNs have initiated communication between the IoT devices. Though IoT offers access to the highest amount of information collected through WSNs, it leads to privacy management problems. Hence, this paper provides a Logistic Regression machine learning with the Elliptical Curve Cryptography technique (LRECC) to establish a secure IoT structure for preventing, detecting, and mitigating threats. This approach uses the Elliptical Curve Cryptography (ECC) algorithm to generate and distribute security keys. ECC algorithm is a light weight key; thus, it minimizes the routing overhead. Furthermore, the Logistic Regression machine learning technique selects the transmitter based on intelligent results. The main application of this approach is smart cities. This approach provides continuing reliable routing paths with small overheads. In addition, route nodes cooperate with IoT, and it handles the resources proficiently and minimizes the 29.95% delay.

WSN is the leading technology necessary for the execution of the IoT structure. IoT’s operational ability and energy established the network communication, cost-effectiveness, dependability, stability, and dynamic function [

Logistic regression (LR) is a type of machine learning algorithm for continuous variables. Though, LR is a classification algorithm, not a constant variable forecasting algorithm [

The ECC algorithm is public key cryptography established on the algebraic structure of elliptic curves through finite fields. ECC is the best algorithm since it is a smaller key size. As a result, it is a power-efficient cryptosystem. The field utilized for the elliptic curve is defined over a prime number [

IoT security is attracting an increasing concentration from both the industry and academic fields. IoT devices are prone to several security attacks and leakage of information. This can potentially offer a broad attack surface for an attacker. IoT devices with weak authentication necessities can be simply compromised and controlled as part of an attack; as the number of connected devices increases, this attack surface continuous to grow. Though, Hybrid Secure Routing and Monitoring (HSRM) with IoT approach using multi-variant tuples using Two-Fish symmetric key approach to determine and avoid the adversary. This approach is preferred to be built through success to the assets of both multipath optimized link-state routing with

To solve this problem, the Secure IoT structure established a WSN using the Machine Learning concept is introduced. This paper objective is to forward the information through intelligent transmitter thus, minimizes the packet losses in the WSN. In addition, the receiver received the secured information from sender to receive by ECC algorithm.

This approach has the following contribution.

The Logistic Regression machine learning technique selects the optimal transmitter based on intelligent results. Logistic Regression machine learning objective is to establish a secure IoT structure for preventing, detecting, and mitigating threats.

The Elliptical Curve Cryptography algorithm is a light weight key, and it distributes the security keys. It minimizes the routing overhead.

This optimal transmitter improves energy efficiency and minimizes the delay in the WSN.

ECC is lengthily utilized in several multifactor authentication approaches. The threat model deliberates several kinds of attacks comprising Man in the Middle (MIM), denial of service, and weak authentication. Countermeasures to decrease or otherwise evade these attacks are advised. An Intrusion detection system is used to prevent the MIM attack. The Intrusion detection system occasionally catechizes nodes one hop away [

ECC is an efficient key because it is a smaller key. Hence, it minimizes unwanted energy utilization. ECC is a faster transmission, rising for a severe operation. Thus enormous injuries are triggered through intrusion attacks. But, it has not improved the routing efficiency. [

An energy-efficient method established public-key cryptography method attains instant authentication, and it avoids Denial of Service attacks [

A Trust-based Formal Model describes the fault recognition procedure and verifies faults lacking simulating and running. This algorithm is introduced to organized detection models. Though, the trust method does not provide better security [

The objective supervises the IoT surroundings sporadically and forwards the information toward the Base Station (BS), applying multi-hop. Then the destination node cooperates with servers over the internet to store the information. The information is forwarded through the untrustworthy wireless transmission links in the existence of malfunctioning nodes. The malfunctioning nodes produce and distribute the fake Route Request (R_{REQ}) packets; as a result, links are congested in the network. In this approach, we consider several IoT-based sensors which can exchange the information and forward the supervised information to the BS by applying intelligent boundaries nodes. These boundary nodes have an extra aptitude for information processing and taking an intelligent result to transmit the information toward servers for storage.

The stored information on servers can be transmitted to the smart devices of the end-users via the internet. The IoT devices are established during the authentication phase to be coordinated with companionable devices that are noticeable as suitable entities for the information routing table. Initially, the BS broadcasts the sensor identities (IDs), boundary nodes Ni in the network. All nodes obtainthese broadcast messages; all IoT nodes keep the edge node’s IDs in their table. In this approach, we use the ECC that forward secure information.

ECC is a public key cryptography algorithm, and this model detects several kinds of attacks comprising man in the middle, denial of service, and weak authentication. The ECC model utilizes two keys: the public key and the private key. The public key encrypts the real information; the private key decrypts the original information. Thus, the attackers can’t read and modify the original information. ECC method provides secure information transmission and improves node Reliability.

The Procedure for LRECC approach is given below.

In the LRECC, the IoT sensor nodes gather the information and forward it to the transmitter node chosen by the regression analysis machine learning technique. This technique provides an intelligent result. The regression analysis is necessary through the information that the election of forwarder nodes established on the transmission time, the packet sent, and the response of the packet through obtaining node. In this process, the regression analysis is aimed to determine the effects of independent variables over the dependent variable.

Because of, the sensor node loss rate in the communication procedure as a dependent variable and forwarding information, along with its instant time as independent variables, any node can forward information packets p_{a} at time t_{a} toward the neighboring node by loss rate LR_{a}. These details are kept in the table during the information forwarding and receiving to perform the regression analysis, applying

Here, LR_{a} indicates the loss rate, _{a} denotes the time instant variable, and

In the same way, the values’ y-intercept

The adjacent node with the minimum packet loss nodes are selected as a transmitter for transmitting the information to the BS. In the next part, the LRECC model forwards the gathered information to the BS via route nodes. These route nodes execute the encryption function, applying the ECC technique. ECC is public-key cryptography, and it contains a public key and a private key. Regard as two sensor nodes are communicated among them, they agree to generate an ECC key R. Assume A and B sensor node’s private keys are

nA and nB correspondingly. sensor nodes A and B public keys are specified below.

If A sensor desires to forward a message m to B, A applies B sensor public key to encrypt the message. The ciphertext is specified below.

Here ‘k’ represents the arbitrary number. The arbitrary k ensures that the ciphertext produced is diverse each time, even for an identical message. This provides a hard time for attackers illegitimately demanding to decrypt the message. B decrypts the message by subtracting the organize of kR multiplied via nB from

Here, nB is the private key of sensor B, and it can decrypt the message of sensor A. Here, the sensor nodes are authentic and regression analysis is executed to choose the optimal transmitter. It guides to low transmission and overhead computing overhead. The integration of route nodes minimizes the communication distance with IoT devices and the consumption of resources at minimum expenses. As a result, regression analysis chooses the node with minimum loss rate as a transmitter. Furthermore, the security of LRECC raises reliability.

In this paper, we use the network simulator ns-2.35 to measure the network performance of the HSRM and LRECC approaches. Here, we use 100 sensor nodes, and these sensor node’s transmission range is 200 m. 10 unreliable sensor nodes are arbitrarily distributed in the field. The parameters applied for the investigation are described in

Parameters | Values |
---|---|

Sensor nodes count | 100 |

Unreliable sensor node count | 10 |

WSN traffic | Constant bit rate |

Node distribution | Arbitrarily way |

Simulation time | 100 s |

Sensor node medium access control | 802.15.4 |

Sensor node queue | Priority queue |

Transmission range | 200 m |

The function of the LRECC is measured by delay, remaining energy, throughput, the ratio of packet losses, and overhead of routing.

From

From

It is clear that the LRECC approach minimizes the routing overhead compared to the LRECC approaches. This approach uses the ECC method for security. It reduces the routing overhead. It is due to the regression-based machine learning technique for transmitter node selection which allows efficient communication between the neighboring nodes. The node chosen procedure is established by the stored information concerning the packet loss rate of every neighboring node.

This paper presents Logistic Regression machine learning with the ECC technique to enhance the performance to prevent, detect, and mitigate threats. This paper objective is to forward the information through intelligent transmitter thus, minimizes the packet losses and receiver received secured information in the WSN. This approach uses the ECC algorithm to generate and distribute security keys. ECC is a light weight key; thus, it minimizes the routing overhead. This cryptography technique verifies the sensor nodes and provides better security in the network. Furthermore, the Logistic Regression machine learning technique selects the optimal transmitter based on intelligent results. This optimal transmitter improves energy efficiency and minimizes the delay in the WSN. An extensive simulation illustrates that the proposed LRECC reaches better throughput and it provides continuing reliable routing paths with small overheads. In addition, route nodes cooperate with IoT, and it minimizes the delay. The future enhancement of this approach is to include node mobility. Furthermore, it will be employed in IoT-established WSNs to secure the application environment, for example, smart cities.