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

24 Dossier | Einride next-gen Pod Asked about the challenges the programme has had to overcome, Hallgren says the biggest obstacle so far has been getting approval for a vehicle such as the Pod to operate on public roads. “Right now, legislation varies by country and even by continent, so we work closely with the Swedish traffic and transport agencies for the necessary permits to operate here, and with other international government bodies, to ensure that new legislation fosters innovation. But progress can often be stifled by bureaucracy,” he notes. He reports that the team has been pleasantly surprised and encouraged by the levels of excitement and support from Einride’s partners, both on the shipping and carrier side and on the technological side. “Transforming road freight must be a collective effort, and requires systemic infrastructure changes beyond our scope, so having the right partners on board to accelerate that change is crucial,” he says. On the autonomy front, he argues that the nature of freight transport is very well suited to SAE Level 4 in that, unlike door- to-door passenger transport for example, most sources and destinations for goods – production sites, warehouses and so on – are known in advance, as are the routes between them. Consequently, Hallgren says, a very large proportion of freight transport needs can be met with Level 4 autonomy. “Since a Level 4 system operates within a well-defined ODD, traditional tools for requirements engineering, verification and validation are straightforward to apply, both in terms of verifying functionality within the ODD and detection of the boundaries of the ODD,” he says. “A rigorous quality assurance process is used to verify and validate requirements at all design levels and to feed the results back into the development.” He adds that Level 4 is achievable now for the AET 1 and 2 ODDs, and is well within reach for AETs 3 and 4. In contrast, SAE Level 5 autonomy is much tougher, because of the complexity involved in proving that no human intervention will be needed. At the moment, a remote operator monitors the Pod continuously, to provide assistance if necessary and to control the vehicle directly during any part of an operation that has to take place outside the level 4 ODD. High-end computing platform The need for an automotive-grade computing platform built for high integrity, safety and performance led Einride to choose Nvidia’s Drive AGX Orin system on chip (SoC). Launched in December 2019, the Orin integrates Nvidia’s next- generation GPU architecture, called Ampere, and ARM Hercules CPU cores, along with new deep learning and computer vision accelerators. It is capable of performing 200 trillion operations per second, which Nvidia says is seven times the performance of its previous-generation Xavier SoC. Nvidia stresses that the Orin is designed to handle multiple applications and deep neural networks running at the same time while meeting systematic safety standards such as ISO 26262 ASIL-D. A software-defined platform, April/May 2021 | Unmanned Systems Technology DeepMap uses high-performance Lidar sensors to build and update 3D maps of autonomous vehicle operating environments (Courtesy of DeepMap) Rooftop view of the Pod showing some of the GNSS navigation and 4/5 G comms antennas as well as a forward facing-camera system at ‘brow’ level and Lidars on the corners