Software boosts autonomous driving by understanding people

A new startup is developing AI software that can analyse and anticipate human reactions for autonomous driving.

One major hurdle for companies developing self-driving vehicles is how to anticipate the behaviour of the driver at any given moment. Is the driver distracted, irritated, nervous? If so, how will this affect their driving and when is the right moment for an autonomous system to take control of the wheel? German start-up Bleenco is focusing on overcoming this problem with an artificial intelligence software package called Pitts, which focuses on the driver. Pitts is a framework for allowing real-time predictive analysis on Internet of Things (IoT) devices, including autonomous vehicles.

Many companies are working on developing products for self-driving cars, including a steering wheel that integrates gesture controls and a system for creating predictive sight. Bleenco, however, is focusing on humans. They are working to develop software which can detect and analyse human emotions. Their AI-based system uses photos, sound recordings and physiological sensors to tell whether the driver is, for instance, irritated or distracted. If it determines the driver is not in a fit state to drive at any given moment, the on-board system can then switch the vehicle to autonomous mode.

Bleenco is also working on creating a holistic toolkit, called Carjs, which will allow companies to more easily build cross-platform IoT apps. Carjs allows synchronization of IoT technology and sensor data, to allow the updating of in-vehicle apps on the fly. Their goal is an ‘intelligent’ autonomous vehicle system that will be able to spot pedestrians, heavy trucks or a slippery road, and adjust vehicle speed accordingly. Although Bleeco’s initial focus is on autonomous driving solutions, they also hope to extend the use of their AI software to other areas. With almost 1.3 million road-related deaths each year, Bleenco are hoping to use their proprietary products to prevent crashes and make driving smoother and easier. What other uses could there be for software that can analyse human emotions and behaviour?