Partnership agreement between Outsight and Hesai

Frequently Asked Questions

• What is the Hesai Pandar128 and what makes it useful for infrastructure deployments?

  The Pandar128 is a mechanical spinning LiDAR from Hesai with 128 laser channels, producing denser vertical resolution than earlier 64-channel designs. More channels mean tighter point spacing at a given distance, which improves reliable detection and classification of objects like pedestrians and smaller vehicles in fixed infrastructure scenarios. The sensor was initially targeted at autonomous driving, but its high channel count makes it equally suited to wide-area infrastructure coverage where vertical resolution determines whether a person is distinguishable from a bollard at range. Outsight, whose SHIFT platform is LiDAR-native and compatible with Hesai hardware, pairs sensors like the Pandar128 with its real-time 3D processing pipeline to build infrastructure-based deployments across airports, train stations, and factories.

• Why does a LiDAR sensor manufacturer like Hesai need a separate software partner rather than shipping its own perception stack?

  Sensor manufacturers optimize for photonics, optics, and signal reliability, which is a distinct engineering discipline from real-time 3D perception algorithms. Writing production-grade detection, classification, and tracking software across diverse deployment environments (airports, factories, road intersections) requires a separate, sustained software R&D effort. Outsight, for example, has spent two decades focused specifically on LiDAR software, resulting in 73 patents and a processing pipeline that operates end-to-end in under 50 milliseconds. Maintaining compatibility across firmware revisions and integrating with downstream operational systems adds further scope that a sensor manufacturer would struggle to absorb without diverting resources from hardware innovation. Splitting these responsibilities lets each company focus on its core competency rather than building a full vertical stack independently.

• What role does a French distributor like Cadden play in an international LiDAR hardware-software partnership?

  A regional distributor handles import logistics, local regulatory compliance, after-sales support, and field relationships with system integrators who may lack direct access to an international manufacturer. Including Cadden in the Outsight-Hesai partnership means French customers and integrators can source Hesai hardware through a local channel while accessing Outsight's LiDAR data processing software through a single coordinated engagement. This structure reduces procurement friction for projects that would otherwise require parallel contracts with two overseas entities. For deployments of the SHIFT platform, where Hesai sensors feed real-time 3D perception pipelines, having a local hardware distributor also simplifies on-site commissioning, spare-parts logistics, and ongoing technical support across the project lifecycle.

• How does adding a new LiDAR sensor family to a hardware-agnostic software platform actually work technically?

  Each sensor manufacturer uses a proprietary binary packet format and UDP stream structure for its raw point cloud output. A software vendor writes a dedicated driver that parses those packets, normalizes the coordinate frame, applies factory calibration corrections, and delivers a standardized point cloud to the processing pipeline. Once that driver layer exists, all upstream algorithms (object detection, tracking, behavior classification) run without modification. The effort is concentrated in the driver, not the perception logic. This is the architecture behind Outsight's multi-vendor compatibility: the SHIFT platform supports hardware from Hesai, RoboSense, Ouster, Velodyne, and Seyond through exactly this driver abstraction, allowing new sensor families to be added incrementally without rewriting the perception stack.

• What is mobile robotics perception, and how is it different from infrastructure-based perception?

  Mobile robotics perception runs on sensors mounted on a moving machine. The sensor itself moves, so the algorithm must continuously re-estimate both the robot's own position and the positions of external objects simultaneously. Infrastructure-based perception uses sensors fixed to ceilings, poles, or gantries; the sensor pose is static and known after calibration, so the algorithm dedicates all compute to tracking the moving objects in the scene rather than solving for ego-motion. Outsight is built around this second model: its Infrastructure-based Physical AI approach processes LiDAR data from fixed infrastructure sensors to generate a real-time Motional Digital Twin of everyone and everything moving through a site. The two approaches share sensor hardware and point-cloud processing primitives but require different algorithm architectures and carry different performance trade-offs.

• What industries beyond automotive were early adopters of 3D LiDAR perception software?

  Aeronautics and factory automation were among the earliest non-automotive adopters of 3D LiDAR perception software. Early customer references included Safran in aeronautics and Faurecia in automotive supply. Airports became a high-growth vertical shortly after, with the Paris-Charles de Gaulle deployment by Groupe ADP representing one of the first large-scale infrastructure use cases, now extended across facilities like Dallas Fort Worth and Rome Fiumicino through Outsight's SHIFT platform. These verticals share a common requirement: real-time spatial awareness in dense, safety-sensitive environments where camera-based analytics face lighting, occlusion, or privacy constraints that LiDAR avoids structurally. Factory deployments at BMW followed a similar logic, prioritizing anonymous, precise motion tracking on the production floor.