Issue link: https://www.e-digitaleditions.com/i/1132978
12 I ICT TODAY It will not be easy. Several leading global IP camera manufacturers develop their own SoCs, retaining a tighter hold on compatibility and profit. The designer's challenge is to specify the most cost-effective devices running AI and to use different sensors effectively, lowering the overall project cost through differen- tiation and increasing the reliability of detecting perimeter breaches. SECURITY INDUSTRY INFLUENCERS Some of the most promising and impactful technologies are being driven by industries other than security. This is a well-established technology introduction cycle, even with security's conservative adoption speed. The HEVC (h.265) video compression originated partly from the broadcast and image recognition industries and is now included in most new IP camera product introductions. 802.3bt The ICT industry is powering cabling efficiencies for the security industry and its devices, such as wireless transmitters requiring higher current through the IEEE 802.3bt standard. Security integration professionals may not realize that a high throughput wireless multiple input, multiple output (MIMO) transceiver powered by 802.3bt compliant infrastructure will power one or even FIGURE 3: The new breeds in AI chipsets and algorithms have led to advances in video, including fluid vehicle motion. CLOUD OR EDGE IN THE AGE OF AI? As fast as cloud computing and cloud-based AI has centralized processing, IoT devices create a better and more automated security net. IHS Markit fore- casts 125 billion such devices by 2030. According to Accenture's 2018 Technology Vision Report, "To fully enable real-time intelligence, businesses must shift event-driven analysis and decision processing closer to points of interaction and data generation. Delivering intelligence in the physical world means moving closer to the edge of networks." Moving AI technology to the edge may be leveraged for a variety of security and safety applications moving forward; however, thermal management of the micro- processor has been the limiting factor. The new breed of AI chipsets runs complex AI algorithms more efficiently. Instead of seeing a choppy video feed of a multi-lane freeway, fluid vehicle motion is observed; the make, model and vehicle license plates are captured and even the recognition algorithm's confidence level is usually displayed (Figure 3). This silicon is already in several IP cameras from leading manufacturers and even in a popular video doorbell device, but the BICSI RCDD will need to do some research to find these more capable devices. several lower power PoE IP video cameras, LiDAR devices or IP-based acoustic sensors. Continuing education at BICSI conferences is critical, such as learning about 802.3bt, zone cabling, and thermal impact on cable length due to bundling. These cost-saving techniques are second nature to BICSI professionals; however, many security projects get deployed without leveraging these technologies.