Sensors for the car of the future: Depending on a vehicle’s category and how it’s equipped, it will have between one and three radar systems. Soon up to five will be offered, together ensuring an ”all-around view” that makes new functions like the intersection assistant and parking assistant possible. In the autonomous vehicle, which the automobile industry expects to be available from around 2020, at least ten radar systems may be installed.
The market leadership of Infineon in the rapidly growing market for radar chips for driver assistance systems was confirmed by market research company IHS Technology. While Infineon has sold a total of 20 million radar chips in the past few years, the company intends to have shipped a further 30 million chips for driver assistance systems this year alone. Infineon would have thus doubled its radar chip sales annually for five years in a row.
Texas Instruments: Automation would be the Key driving trend leading to implementation of sensors, right from low end position sensors to high end image sensors in Smart Factories. We see end point-of-capture processing as the key change to help implement a distributed computing eco-system, enabling lesser over-heads and lot reliable implementation. The need to connect all nodes in a typical Factory environment with minimal Wiring running around would also lead to the need for Wirelessly connected end nodes including Sensors, Gateways and control units.
ON Semiconductor: The evolution of image sensor technology is driven at its core by the needs of applications. While these can vary across different markets, a number of commonalities can be found in industrial applications – the ongoing trend toward faster output speeds, higher resolutions, lower noise, improved light sensitivity, smaller optical formats, etc. Unfortunately, these performance vectors are often at odds with each other – while smaller pixels provide higher resolutions and/or smaller optical formats, by themselves they also tend to reduce signal to noise. The challenge, then, is to combine different technologies – whether new pixel geometries, unique color filter array patterns, backside thinning, stacked pixel architectures, or more – in ways that provide improved performance in one or more vectors without sacrificing performance in others. Ultimately this is our goal – not to deploy new technologies as an end unto itself, but to provide products that offer new solutions to the image problems being faced by customers.
Sasken: Making more efficient or quicker edge analytics. The tiny sensor that could be mounted/deployed inside heavy equipment. This would help user to get the inside picture of equipment along with other data from sensors which makes decision making more accurate. The operating range of sensors and product needs to be enhanced to suit different Industrial environment. Integrating more sensors along with image sensor to take out few transmitters on site to monitor process and reduce the cost of automation. Combination of centralised and distributed image sensors and algorithms – The specific task could be done at node level and do more AI based computation in centralised node to build predictive maintenance capability. Battery operated, power efficient cameras which could be deployed with some wireless connectivity to give flexibility to user. Interoperability, connectivity and standards are going to be key things in the area of machine vision. Artificial intelligence adaptation to solve specific problems is a key to manufacturing and process industry. Image sensor and algorithm used for monitoring in current focus in Industry but may be used for controlling in future but may not be near future.
