Part 1 – Medium Access Control

This part will assess your knowledge and skills in modelling an area over which things are distributed and building a MAC protocol.

1. With the help of the performance estimator, generate 3 instances of an Internet of Things (IoT) scenario and comment on the relationship between power and performance of the network. [40%]
   Explain in detail:

   1. (i)  The different parameters that characterize the network.

   2. (ii)  The latency experienced at different hops.

   3. (iii)  Battery vs number of hops.

   4. (iv)  Transmission frequency vs consumption.

   5. (v)  Latency vs consumption.

   6. (vi)  Network build time vs consumption.

2. With the help of Python simulate Slotted Aloha for your instances. [40%]

   Analyze in detail (use plots and text for presenting your results):

   1. (i)  The impact on slotted Aloha‘s performance with different packet arrival rates.

   2. (ii)  Adjust the slot size and assess how it affects the protocol’s performance.

   3. (iii)  Explore the consequence of collisions and discuss potential collision handling strategies.

3. Now consider a multi-user scenario with N = 10 sources transmitting based on the Slotted-Aloha protocol. Assume that each source has a packet generation rate of λ=〖10〗^3 packets per second and packet length 〖M=10〗^3 bits, transmitting over a channel with transmission rate R. Given that a normalized traffic L = 1 is experienced, evaluate the transmission rate. What is the impact of a higher packet generation rate? [20%]
   
   

   Part 2 – Physical Layer

   The company is looking to acquire a new wireless technology. You have been told that the inventors are using 16-QAM as a modulation scheme in the 2.4GHz band. It claims to offer bit rates of 100kbit/s, 50kbit/s, 1Mbit/s and 5Mbit/s at a BER = 10-4 and that the system will work up to 2km.

   You have been asked to prepare a document evaluating the technology. The company has asked for the following issues to be considered in your document (maximum six pages):

   1. Create the QAM and OFDM objects and generate the transmit signal. [30%]

   2. Characterize the impulse response of the channel. [30%]

   3. Provide a short summary of the regulatory conditions in the UK associated with this frequency band. [20%]

   4. Provide a short summary of competitor technologies. [20%]
      
      

      Part 3 – Network and Transport

      In this part, you are going to produce a design for an IP network of a commercial ISP in the UK. The ISP has about 30% of the market of Russian commercial residential broadband uniformly around the country and provides to its users a 100 Mbit/s connection. You should include:

      1. The backbone topology of the network with realistic IP addresses and OSPF weights. [25%]

      2. Proposed capacity of each link taking into account population in each region/city. [25%]

      3. Using Wireshark imagine you are capturing a portion of the traffic in one of the routers in your network. Use Python to analyze the performances of the network and the factors that influence its functionality. [50%]
         
         

         Part 4 – Data Analytics

         Consider an AI-assisted scenario centred around predictive maintenance for water pumps. In this context, you've been given access to a dataset (sensor.csv) containing information from 52 distinct sensors, along with timestamps and the water pump's status.

         1. Data Visualization: [20%]
         (i) Create visualisations illustrating the variation of each sensor's value over time.
         (ii) Generate a count plot displaying the quantity of the unique labels of the machine status.

         What insights can you derive from the histogram?

         2. Data Exploration: [40%]
         (i) Plot the Pearson correlation of the data with a correlation coefficient greater than 0.9.

         What insights we can derive based on the produced results and task A.a? Is it possible to group any of the sensor data together? If yes, could you provide an example of such a group?

         3.

         (ii) Produce a table containing descriptive statistics, summarizing the central tendency, dispersion and shape of a dataset’s distribution, for the sensor data.
         (iii) Compute the duration, in terms of the number of days, for which the data was collected.

         Data Pre-processing: [40%]

         (i) Identify and count the number of null values per attribute, then remove entries with null values.

         (ii) Identify and count any duplicated entries and remove them from the dataset.
         (iii) Encode the data in the machine status column.
         

         (iv) Determine the data types of the sensor data, and normalise the relevant input features.