Autonomous vehicles detect their surroundings, steer, brake, and accelerate – but the real core of autonomy lies in between: in the software that plans, predicts, and decides. This is precisely where driving behavior is created from data. And this is precisely where it is decided whether a system is truly autonomous – or merely reactive.
Each control cycle involves a complex chain of events: the environment is scanned, objects are classified, and the vehicle's own location is determined. The system then predicts the behavior of other road users, calculates a safe path, and implements it immediately via drive-by-wire. This process is repeated every ten to 50 milliseconds – faster than the blink of an eye.
As Prof. J. Christian Gerdes from the Center for Automotive Research at Stanford University emphasizes: "Driving is not just about recognizing – it's about reacting. In autonomy, the delay between input and action can mean the difference between safety and accident."
Software stack: Think in layers, act redundantly
A modern autonomous system consists of modular software components, each of which fulfills a specific function – and together enable dynamic overall behavior. These include:
Decision architecture: Fail-operational is a must
In fields such as defense, port logistics, construction, or mining, failures are not an option – neither at the sensor level nor at the decision-making level. In the event of a fault, systems must not only shut down, but also respond in a controlled and traceable manner.
This is exactly what NX NextMotion is designed for: The platform has four separate control paths with independent decision-making logic, double-secured power supply, and selectively controllable actuators. In combination with safety-oriented watchdog monitoring and built-in status diagnostics, it meets the requirements of ISO 26262, ISO 61508, UNECE R155, and UL 4600 – including the provision of forensically readable black box logging.
Application determines architecture: adaptive intelligence instead of standard solutions
Whether geofencing in public transport, dynamic path adaptation in logistics, or ethically regulated systems in defense applications – every autonomous vehicle has different requirements in terms of the depth and redundancy of its decision-making logic.
NX NextMotion offers an application-open motion layer that reacts deterministically to planning decisions at the system level. In agriculture, it enables robust implementation of pre-planned paths in difficult terrain; in military platforms, it allows a clean separation between decision and execution – even in the event of temporary radio loss or external teleoperation.
Transparency as a technical criterion: Approval requires traceability
Simulation, validation, and auditability are not additional functions, but core requirements of modern autonomy development. Regulators demand scenario coverage (e.g., through PEGASUS or ASAM OpenSCENARIO), explainable decision models, and documented system histories.
NX NextMotion delivers this traceability through an integrated control protocol system that records control pulses, state changes, and reactions in real time. This provides developers, testing agencies, and OEMs with a reliable data basis – both for approval and for continuous safety monitoring during operation.
Conclusion: Autonomy requires reliable execution – and clear system responsibility
Autonomous systems must not only make "smart" decisions – they must also implement these decisions safely and transparently. NX NextMotion was developed at Arnold NextG precisely for this purpose: as a motion-oriented execution layer that can be seamlessly integrated into certified decision-making architectures – fail-operational, modular, and audit-proof.
We control what moves!
Each control cycle involves a complex chain of events: the environment is scanned, objects are classified, and the vehicle's own location is determined. The system then predicts the behavior of other road users, calculates a safe path, and implements it immediately via drive-by-wire. This process is repeated every ten to 50 milliseconds – faster than the blink of an eye.
As Prof. J. Christian Gerdes from the Center for Automotive Research at Stanford University emphasizes: "Driving is not just about recognizing – it's about reacting. In autonomy, the delay between input and action can mean the difference between safety and accident."
Software stack: Think in layers, act redundantly
A modern autonomous system consists of modular software components, each of which fulfills a specific function – and together enable dynamic overall behavior. These include:
- Mapping & localization with SLAM algorithms and HD maps
- Object recognition & prediction via Kalman filters, deep learning, and reinforcement learning
- Path planning using algorithms such as Hybrid A*, RRT*, or behavior-based planning
- Decision logic based on rule-based states, supplemented by machine learning components
- Fallback strategies in case of malfunctions: replanning, emergency stop, or transfer to teleoperation
Decision architecture: Fail-operational is a must
In fields such as defense, port logistics, construction, or mining, failures are not an option – neither at the sensor level nor at the decision-making level. In the event of a fault, systems must not only shut down, but also respond in a controlled and traceable manner.
This is exactly what NX NextMotion is designed for: The platform has four separate control paths with independent decision-making logic, double-secured power supply, and selectively controllable actuators. In combination with safety-oriented watchdog monitoring and built-in status diagnostics, it meets the requirements of ISO 26262, ISO 61508, UNECE R155, and UL 4600 – including the provision of forensically readable black box logging.
Application determines architecture: adaptive intelligence instead of standard solutions
Whether geofencing in public transport, dynamic path adaptation in logistics, or ethically regulated systems in defense applications – every autonomous vehicle has different requirements in terms of the depth and redundancy of its decision-making logic.
NX NextMotion offers an application-open motion layer that reacts deterministically to planning decisions at the system level. In agriculture, it enables robust implementation of pre-planned paths in difficult terrain; in military platforms, it allows a clean separation between decision and execution – even in the event of temporary radio loss or external teleoperation.
Transparency as a technical criterion: Approval requires traceability
Simulation, validation, and auditability are not additional functions, but core requirements of modern autonomy development. Regulators demand scenario coverage (e.g., through PEGASUS or ASAM OpenSCENARIO), explainable decision models, and documented system histories.
NX NextMotion delivers this traceability through an integrated control protocol system that records control pulses, state changes, and reactions in real time. This provides developers, testing agencies, and OEMs with a reliable data basis – both for approval and for continuous safety monitoring during operation.
Conclusion: Autonomy requires reliable execution – and clear system responsibility
Autonomous systems must not only make "smart" decisions – they must also implement these decisions safely and transparently. NX NextMotion was developed at Arnold NextG precisely for this purpose: as a motion-oriented execution layer that can be seamlessly integrated into certified decision-making architectures – fail-operational, modular, and audit-proof.
We control what moves!
