Issue 37 Unmanned Systems Technology April/May 2021 Einride next-gen Pod l Battery technology l Dive Technologies AUV-Kit l UGVs insight l Vanguard EFI/ETC vee twins l Icarus Swarms l Transponders l Sonobot 5 l IDEX 2021 report

29 Einride next-gen Pod | Dossier to understand the traffic rules and how it should behave, it makes it easier for the prediction and planning software to make more decisions ahead of time so that it doesn’t have to do heavy real-time computation on the spot, Luo says. She adds that high-quality sensor hardware is critical for this, as any measurement errors will add up, particularly in tunnels because of the loss of the GNSS signal and IMU drift, which builds up over time between corrections. In this kind of environment, all the information needed to register the location of the vehicle comes from sensor hardware, particularly Lidar, although Luo says DeepMap has developed software techniques to allow the localisation module to compensate for GNSS loss and IMU drift much as possible. The Pod communicates over 5G where it is available, and 4G otherwise, and Einride is working with partner Ericsson to put the systems it needs in place at customer sites around the world. “Ericsson’s technology is responsible for receiving the wireless comms that the Pod transmits, and helps make sure our data reaches its destination as quickly, reliably and securely as possible,” Hallgren says. He stresses that the connections are also designed to guarantee the levels of latency and bandwidth required to support remote assistance from a supervision and control centre, and adds that the Pod will autonomously perform a minimum risk manoeuvre if its onboard systems detect a loss of comms. More broadly, this is part of the ambitious goal of creating the 5G infrastructure necessary for autonomous road freight at scale. Einride is also exploring V2X systems for comms with the environment around the Pod, including road infrastructure such as smart signs, and other connected vehicles. Cooperative sensing Together, the sensor suite and comms system are central to a perception system for managing multiple interconnected vehicles invented by Par Degerman, formerly Einride’s CTO and now an adviser to the company. Known as cooperative sensing, the idea is to enhance the accuracy, redundancy of information sources and reliability in locating obstacles for the Pod and connected vehicles around it through the use of relative positioning, while also potentially reducing the complexity and, as a consequence, the cost of onboard sensor suites. With existing technology, the company says, autonomous vehicles are capable of driving safely among other vehicles on relatively open roads such as motorways, but challenges arise on urban roads because the more crowded conditions can lead to errors in interpreting sensor information. The conventional way to tackle that is to use a redundant system that makes decisions based on information from several independent sources, but this increases cost and complexity. “It is well-known that a more complex system, consisting of a greater number of components, is more prone to degraded operation owing to a malfunction of any of the components,” according to Degerman. In a patent filed in the US in December 2018, Einride describes a cooperative sensing control unit that embodies the technology it is developing. Although the patent does not refer to the Pod specifically, the unit is designed to retrieve the Pod’s own heading and/ or geographical coordinates along with the relative position of an object or obstacle (first measurement) detected by the Pod’s suite of sensors, and also to receive equivalent data including the relative position of the same object from another vehicle. The unit then retrieves the relative position of the second vehicle as measured by the Pod’s own sensor suite. It then translates the external data into a local coordinate system and combines them with measurements from the Pod’s own sensors. The translated external measurement might, for example, say that an Unmanned Systems Technology | April/May 2021 The Pod is a three-axle truck with a gross vehicle weight of 26 t. Note the low-level sensors absent from later configurations