Repository for SIH 2020 - Problem Code : CK108
DTE OF IT & CYBER SECURITY, DRDO
SECURITY & SURVEILLANCE
CLUSTER CHAMPS
Clustering of air objects based on its trajectory : By DRDO
The position of an object in the air can be indicated by latitude, longitude and altitude for a given time. A trajectory is a stream of such quadruples i (time, latitude, longitude and altitude). Given a large set of such trajectories, without any other information, problem is to cluster them into i meaningful objects such as Helicopter, Fighter/civilian Aircraft, UAV, Cruise Missile, dropped bomb, etc. An optimal scalable solution is desired using i open source tools. Design a system to estimate location of flying object based on its trajectory, provide guidance to missile to shoot them depending on their location when missile will meet the object on its trajectory.
We are making a web application based project which will show the clustered air object and predict its trajectory on the basis of data based on previous trajectories. The technology, we are dealing with is machine learning and neural network. Also, flask and mySQL connectivity is used in backend.
Air object intention recognition is the process of analysing the information of target to explain and judge the purpose and plan of the enemy. It is based on further clustering and sample expansion
Convolutional layers are deployed to extract feature representations from high-dimensional weather features
The model, used to predict the trajectory of the object integrates two modules :
- Anencoder LSTM to embed flight plans into a fixed size feature vector and
- Adecoder LSTM that maps the fixed size feature vector to target the flight trajectory sequence.
Dependencies include access to the classified data such as data of all the flying object in INDIA (latitude, longitude, altitude and time).
The dataset for training a model has not been provided in the description of the problem statement so we are unable to proceed forward. As of the theoretical part we are ready.
- SURYANSH TRIVEDI
- SANSKRITI TIWARI
- MUNEES KHAN
- ROHIT PATHAK
- SHUBHI VIJAYVERGIYA
- SANKALP PATEL