Bosch renews commitment to AI
Stefan Hartung, the member of Bosch’s board of management responsible for energy and building technology, recently stated that "in 10 years, every electronic Bosch product will either utilize AI itself or will have been developed and manufactured with its help.”
According to Bosch, the company will increase its investment in AI, and has earmarked 300 million euros ($456M) to expand the Bosch Center for Artificial Intelligence (BCAI) by 2021. The BCAI currently employs 170 experts; Bosch anticipates that number will increase to 400.
“They are currently working on some 80 development projects — from automated driving to applications in medicine and manufacturing. Many of these are being carried out in collaboration with academic partners, including the universities of Tübingen, Stuttgart, and Amsterdam,” the company said in a recent press release.
One example of Bosch technology that utilizes AI is a camera that uses smart image analysis to locate fires — spotting smoke or flames within seconds, according to the company, which is faster than conventional smoke detector technology.
Hartung will step down as head of the Energy and Building Technology business sector in 2019. Christian Fischer, who joined in Bosch board in October, will take over his role.
New OS for cameras
Bosch also recently established a new company called SAST (Security and Safety Things), which is working on an IoT platform for security cameras that can run on the edge.
Hartmut Schaper, the CEO of SAST, told SP&T News in a September interview, that the company’s operating system, currently under development, will initially be provided to the members of the new Open Security and Safety Alliance, and will soon be available more widely. (The alliance formed in Sept. 2018 with founding members Bosch Building Technologies, Hanwha Techwin, Milestone Systems, Pelco by Schneider Electric and VIVOTEK.)
Ultimately, the OS, which is based on the Android Open Source Project, will be an ecosystem “where people will develop apps for surveillance cameras that haven’t even been thought of today,” he said.
The OS could provide more flexibility and freedom, he added, since developers wouldn’t have to develop for specific cameras but create an app once and have it run on many different cameras.
“We think this is a significant change to the industry, driven by the increasing connectivity, but also by the much higher computing power on the edge,” he said.
The OS has already been tested with facial recognition and licence plate recognition apps as a starting point, said Schaper, and has been well received by integrators who have previewed the technology.