• Developing a smart decision-making solution for ABB Robots using Machine/Deep Learning approach (Deep reinforcement learning).• Developed OpenAI gym compatible custom simulation TestBed/TestBench in python from scratch for ABB FlexPicker (DeltaRobot) using PyGame Library• The simulation TestBed/TestBench can be used to develop custom algorithms or use existing ML library such as Stable Baselines, Keras-RL.• Developed a custom Q-learning algorithm (Model Based) to learn to strategically Pick and Place.• Developed a DQN model for the control of conveyor speed for maximizing place and reducing the robot down time.• Utilized DQN from stable baselines to learn predictive Pick and Place with just raw Images from camera as input (Model Free Learning).• Modelled a Cost Function which calculates the cost for each item/container in real time depending on which the robot does Pick and Place operation with the end Goal of Minimizing the total number of semi-filled output container boxes.• Modelled an algorithm which emulates the actual FlexPicker Pick and Place strategy.• Structured a flexible Observation Space, Action Space and Reward Policy that could be easily modified with change in Goal of Learning.• Built an Interactive Tool for the user to define the Pick and Place strategy, which internally calculates the Reward Policy for the user-defined strategy.• Libraries Used: PyGame for Simulation; MatPlotLib, Seaborn for data visualization, OpenCV, PIL for Image preprocessing for Model free Network; Keras, TensorFlow for backend, Numpy for array manipulations; Keras-RL, Baselines for Model Training.