The new study, published in the International Journal of Intelligent Robotics and Applications, addresses the distributed formation control problem for nonholonomic multi-mobile robots, abbreviated NMMRs, operating in obstacle-rich environments. Led by Ziyu Xu, Dongdong Wang, Yunhui Li, Wei Zhang, and corresponding author Zhonghua Miao, with affiliations at Shanghai University and Henan University of Science and Technology, the work develops a dynamic event-triggered control, or ETC, strategy under a leader-follower structure. Rather than sampling and broadcasting control updates at a fixed clock rate, each robot evaluates a triggering condition and communicates only when the condition is satisfied. The result, demonstrated in both comparative simulations and real-platform experiments, is a substantial reduction in communication burden with essentially no loss in tracking accuracy.

The core of the innovation lies in how the triggering mechanism is designed. In conventional periodic control, every robot in the team updates its neighbors at every sampling instant, regardless of whether anything meaningful has changed. This brute-force synchronization wastes bandwidth, clogs wireless channels, and drains onboard batteries, problems that grow quickly as the number of robots increases. Static event-triggered schemes were previously proposed as a remedy, updating only when the state error exceeds a fixed threshold. But the Shanghai-led team went a step further by introducing a dynamic triggering mechanism, in which the threshold itself is governed by an auxiliary internal variable that evolves over time. This dynamic buffer absorbs fluctuations in the error signal and, crucially, enlarges the inter-event intervals, meaning robots can remain silent for longer stretches between updates. The mechanism is woven directly into the kinematic controller by exploiting the very nonholonomic constraints that make wheeled robots awkward in the first place, turning a modeling nuisance into a design resource.

Formations, however, are only half of the story. The second layer of the architecture handles safety. The researchers incorporated artificial potential functions into an adaptive torque controller through a backstepping-based design paired with a sliding-mode error formulation. In this scheme, attractive potentials pull each robot toward its desired position within the geometric formation relative to the leader, while repulsive potentials activate near neighboring robots and obstacles, generating corrective forces and torques that steer the machine away before contact can occur. Backstepping, a recursive design technique, allows the team to bridge the gap between the kinematic level, where desired velocities are computed, and the dynamic level, where actual motor torques must be produced, and the sliding-mode formulation provides robustness against model uncertainty and disturbances that inevitably plague real hardware. Adaptive elements in the torque controller estimate unknown parameters on the fly, so the robots do not need a perfect dynamic model of themselves to behave correctly.

The combination is mathematically demanding, because stability must be guaranteed not only for the continuous-time closed-loop system but also for the aperiodic, event-triggered updates that punctuate it. The authors report that their framework maintains Lyapunov-based stability guarantees despite the irregular communication pattern, ensuring that the formation errors converge while collisions between robots, and between robots and obstacles, are avoided. This is the central theoretical achievement: proving that speaking less often does not break the contract of safety and convergence that formation flying depends on.

To test whether the theory survives contact with reality, the team carried out comparative simulations against both periodic control and static event-triggered schemes. The metrics of interest were the number of communication updates, or triggering events, and the resulting tracking accuracy of the formation. The dynamic ETC scheme significantly reduced communication updates compared with both baselines while maintaining comparable tracking accuracy, a result with immediate practical weight. In multi-robot systems, the communication channel is a shared, finite resource; every unnecessary message is one that competes with telemetry, perception data, and other teammates. Cutting the message count translates directly into larger scalable swarms, lower latency for the messages that do matter, and longer mission endurance for battery-powered platforms.

The simulations were followed by experiments on a real robot platform, where the proposed formation controller was asked to drive a team of nonholonomic mobile robots through an obstacle-populated workspace while holding a prescribed geometric pattern. The hardware trials demonstrated stability, adaptability, and practical effectiveness, confirming that the torque-level adaptive and sliding-mode components behave as designed when wheel slip, actuator dynamics, and sensor noise enter the picture. The full study includes thirteen figures documenting the trajectory evolution, event-triggering instants, formation error convergence, and obstacle avoidance maneuvers observed in the experimental campaign.

The applications targeted by this line of research extend well beyond the laboratory. Multi-robot systems are increasingly deployed in agriculture, where fleets of machines must coordinate planting, weeding, and harvesting across fields dotted with trees, posts, and irrigation equipment. They are central to warehouse logistics, where squads of mobile shelving units weave between human workers and static racks. They also underpin search-and-rescue operations, environmental monitoring, and coordinated transport of large objects, scenarios in which a dropped communication packet or a mid-formation collision can derail an entire mission. Previous efforts, including graph-based motion planners and null-space-based behavioral controllers, have addressed pieces of this puzzle, but few have simultaneously offered bandwidth efficiency, formal stability, and collision avoidance for robots with nonholonomic constraints.

The leader-follower structure adopted by the team also deserves attention. In this paradigm, one robot, the leader, follows a reference trajectory or navigates autonomously, while the followers track transformed versions of the leader’s state to realize the desired formation shape. This distributed architecture avoids the single-point-of-failure risk of a central planner and scales naturally, since each follower only needs information from the leader and its designated neighbors. The dynamic event-triggering mechanism is applied at the kinematic layer of this structure, governing when each follower refreshes its view of the leader and its peers, while the adaptive torque controller at the dynamic layer runs continuously to execute the latest commanded velocities with precision.

What makes the dynamic, as opposed to static, trigger so effective can be understood through the internal variable that accompanies it. In static ETC, the triggering threshold is a fixed constant, and conservative tuning is required to guarantee stability, which forces frequent updates. The dynamic mechanism introduces an extra state, sometimes likened to an energy reservoir, that grows between events and relaxes the threshold when the system is doing well. When tracking errors are small and the formation is healthy, the reservoir fills and robots happily delay communication. When a disturbance strikes, the reservoir drains, the threshold tightens, and updates become more frequent exactly when they are needed. The closed loop thus allocates its own communication budget adaptively, which is precisely the behavior one wants in uncertain, cluttered environments.

The work arrives amid a broader surge of interest in event-triggered and self-triggered control for multiagent systems, a field that has produced surveys and specialized results spanning consensus, formation tracking, and resilient control under cyberattacks. What distinguishes the new contribution is the union of that communication-efficiency machinery with the specific and thorny dynamics of nonholonomic platforms, together with a fully integrated obstacle-avoidance layer implemented at the torque level rather than as an afterthought in the planner. The authors also ground the work in a lineage of prior results on adaptive consensus and event-triggered integral formation control for networked nonholonomic robots, extending those efforts with the dynamic trigger and the potential-function-based safety guarantees.

Funding for the research came from the National Natural Science Foundation of China under grants 52375107 and 32401712, the Shanghai Agriculture Applied-Technology Development Program under grant T20252014, and the China National Postdoctoral Program for Innovative Talents under grant BX20240217, an agricultural tilt that hints at near-term field trials in orchards and farm environments where multi-constraint navigation is already being explored by overlapping groups.

For the robotics community, the message of the study is that communication efficiency and control performance need not be traded against each other. By letting each robot decide, moment by moment, whether its latest state estimate is good enough to act upon, the dynamic event-triggered framework achieves the same geometric fidelity as a chatty, clock-driven fleet while transmitting far fewer messages, and it does so while threading the fleet safely between obstacles. As multi-robot deployments multiply across farms, factories, and hazardous environments, architectures of this kind may become the quiet backbone that lets large teams move as one without ever raising their voices on the network. The code supporting the findings is available from the corresponding authors upon reasonable request, offering other laboratories a concrete starting point for building on the approach.

Social Mini-Games, or SMGs, are the deceptively simple moments of friction that plague robot deployments everywhere. Two delivery robots converging on the same doorway, a fleet of mobile robots meeting head-on in a narrow hallway, several autonomous vehicles arriving simultaneously at an unregulated intersection: these are the scenarios the survey formalizes. The authors define an SMG precisely as a situation in which the preferred or optimal trajectories of two or more agents collide within a common time interval, typically lasting long enough that someone must yield. What distinguishes these encounters from ordinary navigation is what the researchers term “agency.” In SMGs, small actions carry outsized consequences: unless at least one agent alters its course, a deadlock or collision becomes inevitable. The paper contrasts this with general multi-robot navigation, where agents operate with low to medium agency and their trajectories remain largely decoupled. Crucially, the authors note that humans resolve these conflicts gracefully, modulating their velocity by just enough to let another person pass through a doorway first while barely deviating from their preferred speed, a behavior that remains remarkably difficult for robots to replicate.

The mathematical machinery underpinning the survey is substantial. The authors model multi-robot navigation as a partially observable stochastic game, in which each robot maintains a state combining visible parameters, such as position and velocity, with hidden internal parameters like preferred speed and heading. Robots generate local observations through their sensors, follow control-affine dynamics, and incur running costs based on goal distance, control effort, and deviation from preferred paths. Collisions are formalized through the overlap of convex hulls inflated by safety margins. To quantify when a general navigation scenario tips into an SMG, the researchers introduce a geometric “liveness function” that measures the angle between relative displacement and relative velocity vectors. When two agents approach a capacity-one resource like a doorway with nearly parallel position and velocity vectors, the liveness value approaches zero and perfect symmetry, and thus contention, ensues. Without an explicit symmetry-breaking mechanism, such interactions converge to undesirable equilibria: deadlocks or oscillatory behavior in which neither robot makes progress.

Beyond formal definitions, the survey catalogs a canonical set of SMG scenarios drawn from real-world topologies: doorways, intersections, two-way hallways, L-corners, blind corners, and crowded, parallel, perpendicular, and circular traffic patterns. Each scenario activates a distinct subset of SMG characteristics, exposing different deadlock mechanisms, visibility challenges, and coordination demands. Alongside these scenarios, the paper proposes a battery of evaluation metrics that go well beyond the standard measures of velocity smoothness, delay, and path deviation. Because SMG agents possess high agency, the authors argue, evaluation must capture game-theoretic outcomes. They introduce flow rate, which quantifies density-normalized throughput through a spatial bottleneck; a fairness metric rooted in game theory that measures the gap between each agent’s realized reward and what it would earn under a fully cooperative optimum, with weighted priorities allowing, for example, an emergency robot to justifiably claim the right of way; and an influence score computed through counterfactual analysis, measuring how much one agent’s presence perturbs the control behavior of all others by re-solving the game with that agent removed.

The centerpiece of the work is a taxonomy that organizes existing SMG solvers across multiple design dimensions. At the paradigm level, the authors sort methods into four broad families. Multi-agent reinforcement learning approaches, including value-based methods like Q-learning, multi-agent deep Q-networks, and policy-based frameworks such as MADDPG, treat an SMG as a localized stochastic subgame that emerges wherever non-stationarity plagues an agent’s learning process. Multi-agent path finding methods, including Conflict-Based Search, Increasing Cost Tree Search, and Prioritized Planning, resolve conflicts on discretized graphs but classically assume centralized, fully observable, offline computation. Optimization-based approaches, spanning reciprocal velocity obstacles, model predictive control, and control barrier functions, cast navigation as constrained optimal control in which shared inequality constraints activate among coupled agents. A fourth category collects heuristics, evolutionary methods, and hybrid systems such as genetic-algorithm-based topological optimization for warehouse logistics.

The taxonomy’s finer axes reveal the design trade-offs that define the field. Coordination can be centralized, with a single authority issuing commands; decentralized, with each agent deciding independently from local observations; or distributed, combining centralized training with communication-aware decentralized execution. Communication may be present, allowing agents to share plans and intentions, or absent, forcing reliance on shared norms. Deadlock handling splits into proactive prevention, which plans ahead like a chess player thinking several moves deep, and reactive resolution, which waits for a stall to emerge and then breaks it using rules like the right-hand rule, auctions, or adaptive rotational strategies. The authors also formalize “invasiveness,” a property describing how dramatically a solver perturbs an agent’s preferred trajectory: minimally invasive solvers merely tweak speed without altering course, while invasive ones may force an agent to stop entirely and wait. Cooperation ranges from fully cooperative settings with shared cost functions, exemplified by methods like PRIMAL, to non-cooperative settings where agents pursue private objectives, and observability spans full knowledge of all agents’ states down to minimal awareness in which other robots are treated as static obstacles.

To make these abstractions concrete, the survey profiles representative solvers that have become de facto baselines. CADRL uses deep reinforcement learning with LSTM networks to compress variable numbers of nearby agents into fixed-size observations, producing smooth learned collision-avoidance policies, though without formal safety guarantees. The Right-Hand-Rule method enforces clockwise yielding priorities, and can be implemented through safety barrier certificates computed as quadratic programs, but its rigid ordering can violate fairness constraints. Auction-based approaches such as SocialMAPF let agents bid on priority using control barrier functions, balancing efficiency against occasionally longer travel times. IMPC-DR employs infinite-horizon model predictive control over modified buffered Voronoi cells with warning bands, solving a convex optimization at every replanning step and proving that no stable deadlocks can occur under specified conditions, all while relying only on local communication. Perhaps most striking is ORCA-MAPF, a hybrid in which agents normally navigate with the linear-programming-based ORCA velocity scheme but switch to a locally confined, grid-based path-finding solver, agreed upon deterministically without a central controller, whenever a deadlock is detected; in some tests this lifted navigation safety rates from 15 percent to 99 percent.

Underlying all of these approaches are five properties the authors identify as desirable: provable safety, achievable through control barrier functions that guarantee forward invariance of safe sets or through velocity-obstacle frameworks solvable as linear programs; liveness, the guarantee that robots keep making progress and never freeze in place; welfare maximization, which balances individual priorities against collective goals, such as granting an ambulance robot priority at an intersection; social compliance, the ability to respect human norms and expectations; and scalability, which the paper argues should be measured against the size of the active coupling set of interacting agents rather than the total robot population, since many solvers handle large fleets comfortably until a bottleneck triggers a localized, tightly coupled game.

The survey does not shy away from open problems. Visual-input navigation, in which robots must guarantee safety and liveness directly from raw lidar point clouds and camera feeds without expensive state estimation, remains computationally punishing, as evaluating barrier functions and their derivatives in real time strains onboard hardware. Navigation alongside humans raises questions about whether trajectory-forecasting models that excel on sparse crowd datasets transfer to tightly coupled SMG interactions involving passing, weaving, and yielding. And the authors find current simulation environments wanting: most model only single agents in open spaces, and few, among them SocialGym and SEAN 2.0, approach the configurability needed to train socially compliant robots in geometrically constrained settings. To accelerate the field, the team has open-sourced SMGLib, an extensible simulation environment that lets researchers run planners through canonical SMG scenarios, log metrics, and visualize trajectories in real time.

The authors’ recommendations are pointed: the community should prioritize non-invasive techniques that let agents navigate without dramatic deviations, and should favor deadlock prevention over after-the-fact resolution, since avoiding stalls from the start yields smoother and more reliable motion. Drawing an explicit parallel to the influential 2019 survey that unified and streamlined multi-agent path-finding research, Chandra and colleagues express the ambition that their taxonomy will do the same for social mini-games, giving newcomers a map of a once-fractured territory and giving practitioners a common language for matching algorithms to the doorways, hallways, and intersections where robots must now earn the right of way.

The primary objective of this research was to develop a system that allows robots to autonomously identify and grasp target objects even when faced with obstructions. In many scenarios, the presence of clutter complicates simple robot tasks that humans often take for granted. Think of a restaurant kitchen, a workbench filled with tools, or a home space strewn with various items. In such environments, the ability of robots to adapt and make decisions based on their surroundings is crucial for achieving practical applications. The authors delve into various methodologies to enhance robotic perception and decision-making capabilities.

One of the notable methodologies presented in this research involves advanced computer vision techniques that provide robots with the ability to identify objects in a cluttered space. By employing state-of-the-art deep learning algorithms, the robots can analyze visual data and determine the most effective approach for grasping an object. This is a marked improvement over conventional methods, which often require manual programming and predefined movements. Through the integration of machine learning, the robots can learn from past experiences and refine their grasping strategies based on situational feedback.

Another cornerstone of the research is the development of sophisticated manipulation algorithms. The robot’s ability to not only locate but also effectively grab objects is pivotal to its overall functionality. The authors describe how these algorithms employ a combination of force control and trajectory planning to ensure secure and adaptable grasping actions. For instance, the algorithm can adjust the amount of grip force applied to an object based on its shape and material properties. This fine-tuning capability allows robots to handle a diverse range of objects, from delicate glassware to cumbersome tools, without incurring damage or losing efficiency.

The experimental setup presented by Lachhiramka and colleagues is equally impressive. They conducted extensive trials in both simulated and real-world cluttered environments, showcasing the versatility and robustness of their proposed methods. The robots demonstrated a significant improvement in success rates compared to previous robotic systems, illustrating not only their ability to find and grasp objects but also to adapt to unpredicted scenarios. The results of these experiments underscore the viability of using autonomous robots in practical settings—whether in warehouses, homes, or service industries.

Moreover, the research provides valuable insights into the multi-modal training processes employed to enhance the robots’ capabilities. Utilizing reinforcement learning, the robots undergo a series of training sessions in which they receive feedback based on their performance. This iterative process allows for continuous improvement, leading to increasingly refined grasping techniques. The findings indicate that the implementation of such training methodologies contributes significantly to the robots’ ability to operate efficiently in chaotic environments.

An intriguing aspect of this study is its exploration of the interaction between perception and manipulation. The authors argue that these two functions should not be viewed in isolation. Instead, an integrated approach can make robots significantly more effective. Through tight coupling of sensory data and motor commands, a feedback loop is established that improves the robot’s responsiveness to changing conditions. This synergy between perception and action allows for a more cohesive strategy when navigating through clutter, enhancing overall task performance.

The implications of Lachhiramka’s research extend far beyond laboratory experimentation. As industries increasingly adopt automation to streamline operations, the ability of robots to autonomously interact with their environments will become crucial. From logistics and manufacturing to telemedicine and domestic assistance, the potential applications for such technology are vast. The promise of autonomous robotic manipulation could herald a new era where robots seamlessly integrate into daily human activities, working alongside us to enhance productivity and improve quality of life.

Furthermore, the study’s findings could pave the way for further advancements in the field of human-robot collaboration. As robots become more adept at navigating cluttered spaces, their roles could evolve from simple task execution to more complex collaborative functions. This might include tasks such as assisting elderly individuals with daily chores or supporting warehouse workers in inventory management. By ensuring that robots can recognize and manipulate objects within cluttered environments, we inch closer to a future where technology truly complements human capabilities.

The feedback received from the robotics community has been overwhelmingly positive, with experts hailing the work as a significant contribution to the field. As technology giants and research institutions race to develop the next generation of intelligent robots, studies like this one provide a critical foundation for building more capable, adaptable machines. The advancements in robotic perception and manipulation highlighted by Lachhiramka and his colleagues serve as a beacon of innovation, pointing the way toward smarter, more efficient automated systems.

As we move into the future, the importance of such research cannot be overstated. The integration of advanced robotics into various sectors necessitates a thoughtful and well-researched approach to ensure safety and efficacy. Lachhiramka’s study stands as a testament to the relentless pursuit of understanding and enhancing how machines can operate autonomously amidst the chaos of the real world. With each new discovery and innovation, we take another step along the path to achieving truly autonomous robots capable of grasping their surroundings just as adeptly as humans can.

In conclusion, the research led by Lachhiramka et al. is a remarkable leap toward the realization of highly functional autonomous robots. By addressing the challenges posed by cluttered environments and improving object manipulation techniques, this study lays the groundwork for a future where robots can assist effectively in everyday tasks. The combination of advanced perception, sophisticated manipulation algorithms, and intelligent training methodologies demonstrates the research team’s commitment to pushing the boundaries of what is technologically possible. As this field continues to evolve, we can anticipate revolutionary changes in the role of robots in society—an exciting prospect that underscores the importance of ongoing research in autonomous robotics.

Subject of Research: Autonomous robotic manipulation in cluttered environments

Article Title: Autonomous robotic manipulation for grasping a target object in cluttered environments

Article References:

Lachhiramka, S., Pradeep J, Chandaragi, A.A. et al. Autonomous robotic manipulation for grasping a target object in cluttered environments.
Auton Robot 49, 30 (2025). https://doi.org/10.1007/s10514-025-10214-7

Image Credits: AI Generated

DOI: 08 October 2025

Keywords: Autonomous robotics, manipulation, object grasping, cluttered environments, machine learning, computer vision, reinforcement learning.