Figure Q5 shows an electric wheelchair with obstacle avoidance features. This wheelchair consists of 4 DC motor connected to a mecanum wheels, with 2 infrared sensors in front. The a use of mecanum wheels enables the wheelchair to move in any direction including sideways movement. Motor 3 Sensor 2 Motor 2 Motor 1 Motor 4 - Sensor 1 FIGURE Q5 In Figure Q5, sensor 1 and sensor 2 are infra-red sensors that detect the presence of obstacle either on the left or the right side of the wheelchair. As an expert in the field of artificial intelligence, you have been asked to implement an artificial neural network algorithm in this robot to do the following feat: 1) Moving forward: If both sensors does not detect any obstacle, and then turn all motor in forward direction. 2) Moving right sideways: If only sensor 2 detects an obstacle, then turn Motor 2 and Motor 3 forward, while Motor 1 and Motor 4 backward. 3) Moving left sideways: If only sensor 1 detects an obstacle, then turn Motor 1 and Motor 4 forward, while Motor 2 and Motor 3 backward. 4) Moving backward: If both of the sensors detect an obstacle, and then turn all motor in backward direction. ) a) Based on the given information, construct an architectural diagram that depicts your proposed neural-network. (Set the number of hidden layers to 1 and number of neurons in the layer to 2). Label the diagram with relevant information. (8 marks) b) Based on the given information, generate a training example to be used with back- propagation algorithm. (Use obstacle detected = 1, no obstacle = 0, forward = 1, backward =0.) (6 marks) Based on the neural network architecture obtained in Q5a), you are required to train the neural network based on the knowledge you have acquired in the course BEKC 4873. You are tasked to do the following: i Outline the necessary steps in training the neural network based on the chosen algorithm. (2 marks) ii. Perform the neural network training for ONE (1) iteration using the chosen algorithm. (For simplification, all weights of the network is initialized with 0.5, with a learning rate of 0.1) (6 marks) iii. Validate the performance of your network after the weight update. (3 marks) [25 marks