The importance of connectivity in AV teleoperationsDriverless cars are heavily reliant on data, the LIDAR, RADER, and video sensor systems of autonomous vehicles collect loads of data on every second of every drive. While this might be easy for the AI system on the vehicle to process, sending all of this data to a remote operator somewhere at an instantaneous rate will be highly important for teleoperation to work. Human drivers rely on their sense of sight for safety on the road, so much that a driver with a sight impairment may be considered unfit to drive. For someone driving a car remotely, being able to see the car’s surroundings in real-time, with minimal latency, is just as important. For teleoperation to work effectively, you’ll need uninterrupted, low latency, and high-capacity connectivity that captures live data and relays it across to the teleoperator in real-time. It’s like the car is on a perpetual video call with the teleoperator and wherever it goes, it must deliver crisp and clear video streams while relying on whatever public network is available.
The challenge of connectivity for TeleassistanceThis is in fact where the challenge lies. Dropped calls and choppy feeds are a common occurrence in present-day connected devices. With driverless vehicles, an intermittent connection with a teleoperator is not something we’ll want to experience. In addition to network stability, we’ll need network connectivity with a high bandwidth (can send large packets of data within the shortest time possible), high capacity (can handle several vehicles at the same time), widespread accessibility, and of course low-latency. With 5G connectivity just around the corner, one would think we have our answers to these questions already. However, there are some residual challenges when it comes to connectivity for teleoperations even with 5G technology. You can read more about this in a recent article we published called Why one modem is not enough for teleoperations of robots and AVs. Given all of these challenges, we have to come to terms with the fact that network connections are and will continue to be unreliable and achieving stable connectivity for teleassistance will be quite challenging. With cars moving around at high speed over a large geographical area, there will be variance in connectivity and network latency. Some of the likely technological solutions that will be employed to resolve these connectivity issues include:
- Multiple modems: To keep up, teleoperation systems will have to be adapted to establish a solid connection and utilize multiple network modems to their full potential. This will help minimize the issue of packet loss and latency.
- Packet scheduler: since the system will be designed to make use of multiple connectivity channels, a packet scheduler will ensure that the most efficient channel is being used at any given time by analyzing available network signals and other connectivity metrics.
- Dynamic Forward Error Correction (FEC): to cater to the issue of lost packets, a Dynamic Forward Error Correction (FEC) will be needed to reconstruct the lost packet at the receiving end. This will reduce latency and improve the reliability of the feed being received in real-time.
- Dynamic buffering: a dynamic buffering system will ensure the smoothness of video feeds that are being transmitted. The right technology will reduce buffer time quite significantly and vary buffer size accordingly for optimal results.