Abstract: As the engineering tasks become more and more difficult, and the construction environment becomes more and more complex, a single crane can no longer undertake complex and heavy lifting tasks, however, the multi-crane cooperative lifting robot has strong load capacity, large lifting space, diverse lifting actions, which has been widely used in various engineering occasions. In this paper, the dynamic model of multi-crane cooperative lifting manipulator is established based on Lagrange equation. The traditional ant colony algorithm is optimized and improved for path planning of cooperative lifting, considering the probability of falling into local optimal and plenty of unsolvable ants. A three-valued planar grid environment modeling, pheromone step update strategy based on the best-sub-optimal path length and dynamic adjustment function of pheromone volatility factor have been proposed, the simulation results indicates the obvious optimization effect of the improved ant colony algorithm in terms of the best path length, algorithm running time and other performance evaluation indicators. The lifting environment is established with the three-valued planar grid environment modeling, the simulation results of cooperative lifting verifies the dynamics and improved ant colony algorithm，providing guidance value for improving the efficiency and safety of cooperative multi-crane lifting tasks.

Key words: multi-crane cooperative lifting, multi-robot, improved ant colony algorithm, dynamics, cable-driven parallel robot