Electronic Theses and Dissertation

Jaringan Sensor Nirkabel multihop clustering merupakan infrastruktur komunikasi yang banyak digunakan untuk melakukan pengindraan dan pengiriman data dari sensor menuju base station. Namun, pola pengiriman melalui rute yang sama secara berulang dapat memudahkan penyerang melacak lokasi sumber data dan meningkatkan kerentanan jaringan terhadap serangan sinkhole. Serangan tersebut menarik lalu lintas menuju node berbahaya sehingga menyebabkan kehilangan paket, peningkatan keterlambatan, konsumsi energi yang tidak seimbang, serta kepadatan antrean pada node relay. Penelitian sebelumnya umumnya menangani mitigasi serangan, perlindungan lokasi sumber, efisiensi energi, dan pengendalian antrean secara terpisah sehingga belum mampu menjaga keamanan, keragaman, dan keberlanjutan rute secara terpadu. Penelitian ini bertujuan mengembangkan skema keamanan berbasis Location Source Privacy (LSP) pada jaringan sensor nirkabel multihop clustering melalui tiga model yang saling melengkapi. Model pertama berupa secure rerouting berbasis energi dan evaluasi route safety period untuk memulihkan rute yang terdampak serangan sinkhole. Model kedua berupa Adaptive Multi-Level Route Diversion (AMLRD) berbasis phantom routing untuk meningkatkan keragaman jalur dan mengurangi keterprediksian lokasi sumber. Model ketiga berupa Queue-Aware AMLRD yang mempertimbangkan kondisi antrean, utilisasi relay, energi residual, dan biaya komunikasi dalam pemilihan rute. Penelitian menggunakan pendekatan pemodelan dan simulasi dengan MATLAB pada jaringan berukuran 300 × 300 m² yang terdiri atas 100 node sensor, lima cluster, energi awal 2 joule, ukuran paket 4.000 bit, dan waktu pengamatan 2.500 round. Serangan sinkhole diaktifkan pada round ke-250 dengan radius pengaruh 100 m. Evaluasi dilakukan menggunakan metrik throughput, rasio pengiriman paket, kapasitas pengiriman, efisiensi energi, latensi layanan, jumlah node hidup, masa hidup jaringan, keragaman rute, utilisasi relay, panjang antrean, paket tertunda, dan buffer overflow. Hasil 60 simulasi independen pada model secure rerouting menunjukkan bahwa throughput meningkat dari 242 menjadi 371 kbit/s, rasio pengiriman paket meningkat dari sekitar 70% menjadi 99,2%, dan kapasitas pengiriman meningkat dari 61% menjadi 93%. Efisiensi energi meningkat sebesar 38,28%, sedangkan latensi layanan menurun dari 70,37 ms menjadi 40,10 ms. Kematian node pertama juga dapat ditunda dari sekitar round ke-841 menjadi round ke-1.545. Model AMLRD meningkatkan keragaman rute dan mengurangi penggunaan jalur forwarding yang sama secara terus-menerus, sedangkan Queue-Aware AMLRD mampu menghindari relay dengan antrean padat, mendistribusikan lalu lintas secara lebih seimbang, dan mempertahankan kestabilan pengiriman paket. Penelitian ini menegaskan bahwa integrasi pemulihan rute, diversifikasi jalur, dan pengendalian antrean mampu memperkuat LSP, memitigasi serangan sinkhole, serta meningkatkan keandalan dan keberlanjutan jaringan sensor nirkabel multihop clustering.
Kata kunci:
Location Source Privacy; serangan sinkhole; secure rerouting; AMLRD; phantom routing; queue-aware routing.

A multi-hop clustered Wireless Sensor Network (WSN) is a communication infrastructure widely used for sensing and transmitting data from sensor nodes to a base station. However, repeated data transmission through the same routing paths may enable adversaries to trace the data source location and increase the network’s vulnerability to sinkhole attacks. Such attacks attract network traffic toward malicious nodes, resulting in packet loss, increased delay, unbalanced energy consumption, and queue congestion at relay nodes. Previous studies have generally addressed attack mitigation, source location protection, energy efficiency, and queue management separately. Consequently, they have not fully integrated route security, diversity, and sustainability into a single framework. This study aims to develop a Location Source Privacy (LSP)-based security scheme for multi-hop clustered WSNs through three complementary models. The first model employs energy-aware secure rerouting and route safety period evaluation to recover routes affected by sinkhole attacks. The second model introduces Adaptive Multi-Level Route Diversion (AMLRD), which employs phantom routing to increase route diversity and reduce the predictability of the source location. The third model, Queue-Aware AMLRD, incorporates queue conditions, relay utilization, residual energy, and communication costs into the route selection process. This study adopts a modeling and simulation approach using MATLAB. The simulated network covers an area of 300 m × 300 m and consists of 100 sensor nodes, five clusters, an initial energy of 2 joules, a packet size of 4,000 bits, and an observation period of 2,500 simulation rounds. The sinkhole attack is activated at round 250 with an influence radius of 100 m. Performance is evaluated using throughput, packet delivery ratio, delivery capacity, energy efficiency, service latency, number of alive nodes, network lifetime, route diversity, relay utilization, queue length, delayed packets, and buffer overflow. The results of 60 independent simulation runs for the secure rerouting model show that throughput increases from 242 to 371 kbit/s, the packet delivery ratio improves from approximately 70% to 99.2%, and delivery capacity increases from 61% to 93%. Energy efficiency improves by 38.28%, while service latency decreases from 70.37 ms to 40.10 ms. The first node death is also delayed from approximately round 841 to round 1,545. The AMLRD model increases route diversity and reduces the continuous use of the same forwarding paths. Meanwhile, Queue-Aware AMLRD avoids highly congested relay nodes, distributes network traffic more evenly, and maintains stable packet delivery. These findings demonstrate that the integration of route recovery, route diversification, and queue management strengthens LSP protection, mitigates sinkhole attacks, and improves the reliability and sustainability of multi-hop clustered WSNs. Keywords: Location Source Privacy; sinkhole attack; secure rerouting; AMLRD; phantom routing; queue-aware routing.