الفهرس 
AbstractOnly 14 pages are availabe for public view
 |  | from | 112 |  |  |
| | | from | 112 | | |
Abstract
The Objectives of this thesis are as follows:
• Identification of Satellite Constellation communication challenges.
• Survey DTN and non-DTN protocols for LEO Satellite Constellation Networks.
• Study DTN characteristics and general architecture.
• Robust Communication in Challenging Environments: LEO satellites experience frequent handovers, varying link conditions, and intermittent connectivity as they move rapidly across the sky. DTN protocols are designed to handle such challenging and dynamic environments, ensuring robust communication even in scenarios with high latency or temporary disruptions.
• Contact Prediction: Explore methods for predicting satellite contact opportunities accurately to optimize data scheduling and resource utilization within the LEO constellation.
• Inter-Satellite Communication Enhancement: DTNs facilitate efficient inter-satellite communication and data exchange within the LEO constellation. By allowing satellites to store and forward data to other nodes, even if a direct connection is not available, DTN protocols optimize data routing and distribution across the constellation.
• Resilience to Network Disruptions: LEO satellites may experience outages due to atmospheric interference, eclipses, or other factors. DTN protocols enable the network to adapt and recover from such disruptions, maintaining data communication even during temporary connectivity gaps.
• Scalability for Large Constellations: As LEO constellations grow in size and complexity, DTN protocols offer scalability advantages, making it easier to manage and control a large number of satellites and ground stations.
• Simulation of selected DTN protocols in terms of buffer occupancy, delivery ratio, and delay at different levels of noise and interference.
Methodology
1- The model uses a dataset of Starlink LEO satellite constellations operated by SpaceX. It contains 66 uniformly distributed satellites of different altitudes and inclinations. The Simplified General Perturbations (SGP4) propagator is used to predict the progress in the altitude of each satellite at intervals of length one second.
2- 2- The propagator provides the altitude (state vector) of each satellite in (x, y, and Z) coordinates and its velocity components with their time epoch (t).
3- These data are the main entry to the trace file of the simulator. The DTN routing protocol and CC . In addition to the traffic and wireless channel model. The main difference in the presented work is the mobility model, where the nodes’ mobility is deterministic and steady. While the network topology is dynamic. This is typical behavior of LEO satellites in a constellation.
4- In the CC-applied method, forwarding messages of buffer availability are advertised by nodes. The forwarding process is governed by receiving an availability message from the intermediate node that advertises an available buffer capacity.
5- In the suggested technique, Hello packets which are nodes frequently broadcasting to one another. They are used to spread buffer occupancy information (nodes use the same packets to discover each other). The advertised buffer size Ba is determined by subtracting the current Buffer occupancy (Bo) and the overall buffer size Bs multiplied by the Congestion Threshold (TC).
6- The Hello and anti-packet messages, as an efficient mechanism for reducing the message delivery delay, are applied in the model. Every 100ms, DTN nodes send Hello messages to one another in order to announce the buffer status. After 750 seconds, the bundle’s lifetime is scheduled to expire, and all copies of the bundle will be removed. If the sender doesn’t receive a return receipt within the retransmission timeout (1000 seconds), it will re-transmit the bundle. Return receipts serve as anti-packets. Their lifetime is the minimum of retransmission timeout less bundle forwarding time, and bundle lifetime (750 seconds).
7- Message copies are deleted once a copy arrives at its destination to prevent the negative impacts of congestion.
Results and Findings
1- it shows that the Ep requires a larger size of the buffer by about 52% more than B-SaW because of the limited number of copies for bundle compared with Ep where every node will store a copy of a bundle received from a neighbor if it does not have it, it also shows that applying the CC causes an increase in the buffer occupancy in both Ep and B-SaW by around 62%.
2- the DR in power 12.5 is increased in B-SaW for more than EP because of the increased number of copies, and this causes congestion, The usage of con- gestion control improves the DR of all protocols because if buffers are full, new packs are not generated. In the case of power (11.5, 10.5), the DR decreased because of less power.
3- in terms of DR and DP, respectively. the case of Tx power equals 12.5 dBw, DP characteristics are directly related to the BO for the applied protocols, where the increase in size of the BO causes an increase in DP.
4- the OH in Ep is greater than B-SaW by about 40%. When the CC is applied, the OH is increased in Ep by 20% and in B-SaW by 15%.
5- In terms of the four performance metrics at NF= 2, The comparison shows that reducing the NF causes, in general, a small change in the BO, and much higher utilization in the DR accompanied by an increase in the DP and the OH. The lower NF increases the number of available neighbors, hence increasing the network traffic and the number of DP.
Recommendations
1. Predict the position of the satellite in LEO constellations by machine learning algorithms.
2. Using Machine Learning for Satellite Communications Operations.
3. Optimized the Dynamic inter-satellite link that changes at any time.
4. Using AI to make satellites make decisions in case of delay or intermittent communication in the network.