“HIRI FUTURE” Wins First Prize (Level A) at the China Artificial Intelligence Competition

On December 12, the 6th event, themed “Integrating New Ideas, Pooling Wisdom, and Competing for Development,” was held. China Artificial Intelligence Competition The accompanying forum was successfully held in Xiamen. Grounded in national strategies and aligned with industry needs, the forum brought together stakeholders from government, industry, academia, research institutions, and end-users to jointly explore innovative AI technologies and new development pathways. It also unveiled significant achievements, showcased annual technological breakthroughs, and accelerated the implementation of high-quality projects nationwide.

The highly anticipated award ceremony for the 6th China Artificial Intelligence Competition was also held at the main forum. This competition focused on the business needs of national ministries and commissions, industry hotspots, and the practical realities of industrial development. It featured 10 competition tracks, including a large-model adversarial challenge, a call for AI-powered supply-chain case studies, an AIGC video detection contest, and others. The competition attracted nearly a thousand teams to register, with a total of over 1,200 submitted projects. After intense competition, 100 outstanding teams stood out and were awarded either an A-level or B-level certificate.

The project—"Autonomous Intelligent Agent Cluster Collaborative Operation System for Port Lifting Equipment and Its Application Practice"—submitted by "HIRI FUTURE," a brand under Fujian Electronic Port Co., Ltd., has won the first prize (Level A) in the Embodied Intelligence Application Innovation Competition, showcasing the company's leading strength and innovative achievements in the field of port intelligence.

Overview of the Award-Winning Case
“The Autonomous Intelligent Agent Cluster Collaboration System for Port Cranes and Its Practical Applications” is an intelligent system designed specifically for crane operations in port yard scenarios. It leverages embodied intelligence technology to enable equipment to autonomously perceive, make decisions, and execute tasks. The project was independently developed by Fujian Electronic Port Co., Ltd. and, under the “HIRI FUTURE” brand, has built a comprehensive smart port solution covering perception, decision-making, control, and scheduling. This solution has been implemented in over 80 major ports both domestically and internationally, with a cumulative deployment of approximately 3,000 sets of systems. It currently serves core domestic ports such as Tianjin Port, Shanghai Port, Ningbo Port, Xiamen Port, and Hong Kong, as well as regions along the Belt and Road—including Malaysia, Singapore, Thailand, and Brazil—and BRICS countries. The application scope encompasses scenarios such as container terminals and bulk cargo yards. Focusing on transportation and logistics scenarios, the project addresses key challenges in traditional port crane operations, including low efficiency, high safety risks, conflicts in multi-machine collaboration, and heavy manual labor intensity. By leveraging embodied intelligence technology, it achieves automated, collaborative, and intelligent operations for yard cranes, thereby enhancing overall port operational efficiency and safety.

Core Issues and Solution Approaches
This solution aims to address the core issues of traditional port cranes—reliance on manual labor, efficiency bottlenecks, and high safety risks. The proposed approach is to build an integrated embodied intelligence system that combines perception, decision-making, and execution. By leveraging multimodal sensing, intelligent scheduling, and precise control, the RMG can achieve fully automated operations and remote monitoring in complex yard environments.

Tackling Technical Challenges
In the course of advancing this project, we have overcome four core challenges: 1. The challenge of environmental perception: By integrating data augmentation with human-operated data, we have enhanced the model’s robustness in adverse weather conditions and complex scenarios. 2. The challenge of dynamic decision-making: We have introduced a reinforcement learning framework that incorporates human expertise, enabling anthropomorphic intelligent decision-making. 3. The challenge of collaborative scheduling: We have developed a hybrid architecture that combines centralized scheduling with distributed execution, thereby avoiding multi-robot conflicts and deadlocks. 4. The challenge of human-robot collaboration: We have designed an intuitive remote monitoring and takeover system, ensuring smooth and safe human-robot interaction.

Project Innovation
This project innovatively constructs a “domain-wide adaptive agent,” achieving a transformative leap from traditional automation to true intelligence. Its primary innovation lies in the adoption of an AI multimodal fusion perception system. Leveraging deep-learning-based vision technology, this system not only detects the presence of containers but also precisely identifies critical information such as container numbers, twist-lock statuses, and door orientations. Combined with 3D visual positioning, it dramatically enhances perceptual robustness and accuracy in complex operational environments, providing automated systems with a reliable “intelligent eye.” This innovation directly addresses the challenges in environmental perception—specifically, the instability caused by factors such as stacking-induced occlusions and metallic reflections. Building on this foundation, the solution introduces agent-based autonomous decision-making and control. The crane, equipped with an “experienced driver’s” brain, utilizes reinforcement learning algorithms to achieve adaptive anti-sway control and precise positioning. It can also dynamically plan paths and handle non-standard scenarios, such as AGV docking deviations, significantly improving operational efficiency and smoothness. This marks a fundamental shift from the rigid rule-based approaches described in the “difficulties in dynamic decision-making” toward human-like intelligent decision-making.

Next step

HIRI FUTURE will continue to deepen its expertise.

The Integration of Embodied Intelligence Technologies with Port Scenarios

Use technological innovation to solve industry pain points.

Help more ports, both domestic and international.

Achieve a smarter upgrade that is safer and more efficient.