1. Deep Learning based driver monitoring system (activity & object recognition)
In recent years there has been a lot of focus on developing driver monitoring software for integration in passenger cars and other vehicles to facilitate better safety and other functions that improve the user experience. By studying a person’s posture and body movements, intelligent interior vehicle algorithms can draw conclusions about a person’s alertness, attention and focus.

Tomorrow’s cabin sensing features will include detection of passenger position, safety belt status and forgotten objects, as well as enabling multimodal functionality such as deeper AI and mood recognition.So the car is able to seamlessly transfer control of the vehicle to an awake and able driver, call for help in a medical emergency, or offer to play the perfect song for the moment.

Can you think of an interesting use case(s) in this area and implement them?

1. Nexgen Car upgrade and finance option recommendation engine
Recommend the right car model and financing option to an existing car owner in India on or before their financing contract matures.

The objective of the challenge is to figure out the right offer for an existing car customer who has financed his/her car using a financing option so that they upgrade to a newer model. There are two components to making a great offer; the right car model and a timely financing offer. Through data analysis, make an offer at the right time so that customer feels compelled to upgrade to a new one. Factors influencing the model will be current contract status, term, maturity date, credit status, monthly payment, outstanding loan amount, current price of a new car (of different models) and current contract options.

2. Benz for kids!
Kids with their intrinsic energy and behaviour have to be kept engaged in the rear seats during commutes. Many a times parents do that by handing out their mobile phones or other hand-held devices which is certainly not their favourite or first preference. We would be interested to see solutions in this space where kids can get byte sized learning modules with gamification elements to make it intriguing. With several car and geo location data points from the car, there can be an array of opportunities and ideas which feed into entertainment for kids (and adults too).

1. Safe Route & best way forward
How many times have we taken not the shortest route as shown by the navigation because the shortest path has some desolate, dark stretches we want to avoid. Wouldn’t it be nice to have this information provided too? Most navigation solutions have options like shortest distance, fastest route, toll-free and other variants. With urbanisation, both in developed and developing countries and constantly evolving cities and suburbs there are extensive dynamic situations which influence our commute. Introduction of the ‘safest route’ option in navigation based on different factors, e.g., road repairs, accident and SOS data, newsfeed on disruption, based on reviews, (or anything else you can think of) can be a worthwhile option for the users.

2. Transport solution/algorithm for Indian urban needs (e.g. multimodal transport)
India’s transport demand has grown by almost 8 times since 1980 - more than any other Asian economy with thriving auto industry and allied economic growth. However, with this growth comes few health and welfare related challenges which needs to be addressed. One clear imperative for the country’s mobility paradigm is to build safe, clean, convenient and congestion free transportation. We are seeking ideas to optimize travel footprint by reducing congestion caused by passengers and goods flow in the urban context. Data-based models and measures such as intelligent routing and recommendation based on public infrastructure data, smart algorithms that make multi-modal and balanced mobility a more lucrative option than current.

3. Mobility for Specially abled
Mercedes-Benz is devoted to ensure that people with restricted mobility are easily and independently able to take to the roads, or be safely and comfortably carried as passengers. Inclusion of physically disabled, senior citizens, visually impaired, cerebral palsy and others for usage of all modes of mobility. We are looking for digital solutions with which our special consumers can leverage the currently available ecosystem in a better way.

1. Recognition of thin cross section objects changing direction (e.g. a cyclist)
Object detection in self-driving cars is one of the most challenging and important impediments to full autonomy. Self-driving cars need to be able to detect cars and pedestrians to safely navigate their environment. In recent years, state-of-the-art deep learning approaches such as Convolutional Neural Networks (CNNs) have enabled great advances in using camera imagery to detect and classify objects. Recognizing thin cross section objects, which is also continuously changing direction (e.g. a cyclist) is even more challenging. Can you come out with your approach to do the same and demonstrate it?

2. Image understanding for scene labeling based on sensor input
For autonomous driving, vehicles typically have multiple sensors of complementary modalities such as cameras, LiDAR and RADAR. They generate a comprehensive and robust representation of the surrounding. Each sensor modality leverages its specific strengths to extract as much information as possible from the observed scene. Based on the combined sensor data, a detailed environment model is created. This environment model provides the basis for high-level tasks such as object tracking, situation analysis and path planning.

In order to perform these high-level tasks well, it is essential for an autonomous vehicle to not only distinguish between generic obstacles and free-space, but also to obtain a deeper semantic understanding of its surroundings. Detailed semantic information of similar quality has to be extracted independently from each of the sensor modalities to maximize system performance, availability and safety.

Can you come out with your approach which is different from existing ones and can provide better results?