Artificial Intelligence (AI) is becoming more intelligent for automotive features. Automakers and Silicon Valley are all vying for pieces of the automotive AI pie.
NVIDIA offers the Tegra X1 with Deep Learning. Panasonic introduced a pedestrian recognition demonstration at ITS World Congress. Mitsubishi Electric is working on cognitive driver distraction detection. Nissan premiered the IDS concept pregnant Tesla (LEAF). Toyota has partnered with MIT and Stanford to open an AI center. Tesla released its AutoPilot. While Siemens offers radar-based parking space detection.
ABI Research’s Automotive Safety and Autonomous Driving and Smart Transportation Research Services report that AI especially Deep Learning based on neural network computing, parallel processing and unassisted cloud-based crowd learning are driving key innovations in automotive. AI technology is being used for Virtual Assistants, ADAS and Traffic Management.
- Advanced agents or Virtual Assistants know the driver’s preferences and allowing natural language interaction within the vehicle and driving context. Apple’s Siri, Google Now, and Nuance Dragon represent early examples of in-vehicle integration and adaptation of virtual assistants. Microsoft announced intentions to develop an automotive-grade version of Cortana. Nissan’s Intelligent Driving System (IDS) concept includes a virtual assistant.
- Advanced Driver Assistance Systems (ADAS) and driverless vehicles rely ondeep learning–based machine vision for identifying and recognizing pedestrians and vehicle types, as well as interpreting and predicting complex traffic situations.
- AI is important for connected infrastructure. Adaptive Traffic Lights, dynamic pricing for Electronic Toll Collection (ETC) and Road User Charging (RUC), and automated Intelligent Transport Systems (ITS) will be powered by advanced artificial intelligence exceeding the capabilities of human operators at traffic operation centers today.
AI’s Deep Learning harvests the crowd intelligence of millions of vehicles to accelerate the machine learning cycles. Moreover, it also involves including intelligent roadside infrastructure and the data it generates from traffic cameras, road sensors and toll gates. This will ultimately lead to far-reaching convergence between connected driverless vehicles and ITS, resulting in holistic, remotely controlled and automatically reconfiguring closed loop transportation systems with traffic throughput optimization heavily relying on demand-response approaches.