Emergency service vehicles like ambulance, fire, police etc. should respond to emergencies on time. Existing barriers like increased congestion, multiple signalized intersections, queued vehicles, traffic phase timing etc. can prevent emergency vehicles (EVs) achieving desired response times. Existing solutions to route EVs have not been successful because they do not use dynamic traffic parameters. Real time information on increased congestion, halts on road, pedestrian flow, queued vehicles, real and adaptive speed, can be used to properly actuate pre-emption and minimise the impact that EV movement can have on other traffic. Smart cities provide the necessary infrastructure to enable two critical factors in EV routing: real-time traffic data and connectivity. In addition, using autonomous vehicles (AVs) in place of normal emergency service vehicles can have further advantages in terms of safety and adaptability in smart city environments. AVs feature several sensors and connectivity that can help them make real-time decisions. We propose a novel idea of using autonomous emergency vehicles (AEVs) that can meet the critical response time and drive through a complex road network in smart cities efficiently and safely. This is achieved by considering traffic network analogous to real-time systems (RTS) where we use mixed-criticality real-time system (MCRTS) task scheduling to schedule AEVs for meeting response time.