Prasanna Venkatesh on How AI and IoT Are Empowering Smart Cities
AI and IoT are building the future
Innovation Head of Wave4 IoT Lab at HCL Technologies, Prasanna Venkatesh gave an informative keynote about how AI and IoT are empowering smart cities. The highlights of his speech are as follows.
Check out his whole presentation on our website and YouTube channel.
The evolution of AI and IoT
According to Prasanna, in the early 2000s, street lights were normal and manual. 2G was the connectivity technology around that time, though it was not predominantly available across many parts of the world. During that time, there was no AI installed in the city street lights.
From the early 2000s to now, AI hasn’t been applied much to city street lights. However, a lot of lighting-based communication technologies have grown in the actual ecosystem. 3G and 4G technologies have emerged and smart city monitoring and control stations have had a good amount of AI being applied.
Starting from 2022 onward, we are entering the era of thinking machines, or the emergence of AIoT, which turns normal street lights into thinking street lights. 5G now enables AI to be applied to connectivity technologies and city monitoring and control stations are becoming much smarter.
Thinking Street Lights
Prasanna walked us through the thinking street light life cycle, which includes six stages:
- Thinking street light detects the malfunction
- It creates a ticket and sends it to city authorities
- Smart ticketing system assigns support engineers
- Thinking street lights track the issue resolution based on SLA
- Should there be an SLA breach, thinking street light sends an escalation to take care of it
- AI detects the issue is fixed and thinking street light closes the ticket
He then shared the four challenges of thinking street lights:
- Understanding the fault in the light
- SLA management
- Escalation workflow management
- Intelligence management
Prasanna presented how these issues could be solved with AI. Firstly, the data and events in smart street lights are processed by machine learning algorithms. Then, the system is constantly updated with a knowledge base. The SLA knowledge goals, facts produced from the knowledge base, and insights from the ML system create machine reasoning, which then proposes actions for optimization. Eventually, the optimized actions go back to the smart street lights. Edge AI infrastructure is required for this cycle to happen. To manage this infrastructure, AI model deployment, Edge management, SLA management, knowledge management, machine reasoning management, and optimization management are essential.
Challenges of Edge AI
Since 3 trillion IoT devices are expected to come into existence by 2030, there are some challenges with Edge AI that need to be addressed:
- AI models’ complexity and diversity will grow
- Full stack acceleration
- Deploy and manage scale
Thinking City
What if thinking street lights lead to thinking cities?
Prasanna gave some reference architecture for the potential thinking city:
- AI in connectivity
- AI in Edge
- AI in service layer
- AI in service management
- AI in service delivery
“Some of the beauty of the whole structure is we are seeing it is not just enabling the city, but the city intelligence as a service plays a very important role where city can monetize their actual infrastructure and enable various services to be provided for end users there.” — Prasanna Venkatesh
